Background: Preterm birth (birth before 37 weeks of gestation) and its complications are the leading contributors to neonatal and under-5 mortality. The majority of neonatal deaths in Kenya and Uganda occur during the intrapartum and immediate postnatal period. This paper describes our study protocol for implementing and evaluating a package of facility-based interventions to improve care during this critical window. Methods/design: This is a pair-matched, cluster randomized controlled trial across 20 facilities in Eastern Uganda and Western Kenya. The intervention facilities receive four components: (1) strengthening of routine data collection and data use activities; (2) implementation of the WHO Safe Childbirth Checklist modified for preterm birth; (3) PRONTO simulation training and mentoring to strengthen intrapartum and immediate newborn care; and (4) support of quality improvement teams. The control facilities receive both data strengthening and introduction of the modified checklist. The primary outcome for this study is 28-day mortality rate among preterm infants. The denominator will include all live births and fresh stillbirths weighing greater than 1000 g and less than 2500 g; all live births and fresh stillbirths weighing between 2501 and 3000 g with a documented gestational age less than 37 weeks. Discussion: The results of this study will inform interventions to improve personnel and facility capacity to respond to preterm labor and delivery, as well as care for the preterm infant. Trial registration: ClinicalTrials.gov, ID: NCT03112018. Registered on 13 April 2017. Keywords: Preterm birth, Neonatal mortality, Quality improvement, Simulation training, Kenya, Uganda * Correspondence: Dilys.Walker@ucsf.edu Phelgona Otieno and Peter Waiswa contributed equally to this work. Institute for Global Health Sciences, University of California, San Francisco, CA, USA Department of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco, CA, USA Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Otieno et al. Trials (2018) 19:313 Page 2 of 12 Background authors reported an increase of EBPs from 68% to 95% Preterm birth, defined by the World Health over a 6-month period, as well as a reduction in peri- Organization (WHO) as birth before 37 weeks’ gesta- natal mortality from 22 to 13.8 deaths/1000 deliveries tional age (GA), and its subsequent health complica- . A more recent report from the Better Birth Trial in tions, are the leading cause of both neonatal mortality India revealed that despite an increase in use of EBPs, and under-5 child mortality . Globally, an estimated outcomes did not improve , suggesting that the one million newborns die each year due to complica- checklist alone may be insufficient in some contexts. tions of prematurity, and another 0.9 million preterm Data from other countries’ implementation of the Safe survivors suffer from mild to severe neurodevelopment Childbirth Checklist are still forthcoming, with the hy- impairments . Thus, in order to further decrease child pothesis that checklists improve uptake of EBPs by min- and neonatal mortality and morbidity, averting prema- imizing errors of omission or reminding the target turity and helping preterm infants survive and thrive audience to perform critical steps. must be of highest priority. Second, simulation-based training is a technique used In Kenya and Uganda, annual neonatal mortality rates in many fields to immerse participants in a task or set- have slowly decreased over the last decade, but remain ting that simulates “real-world” contexts. In the obstet- high at 23 and 21 per 1000 live births, respectively . rics field, PRONTO International aims to optimize care Preterm birth rates are estimated to be 12.3% in Kenya during birth, through developing and implementing in- and 13.6% in Uganda . Both countries also have esti- novative in situ competency-based trainings that are mated high rates of stillbirth (per the definition of the grounded in highly realistic obstetric and neonatal simu- WHO – a baby born with no signs of life at or after lation training. PRONTO trainings also emphasize team- 28 weeks of gestation) with 21–23 deaths per 1000 total work and communication, and kind and respectful care, births . Stillbirths are often not measured accurately which contribute to empowering teams to identify sys- and may disguise even higher rates of prematurity and/ tem errors to catalyze change in their facilities. Increased or neonatal mortality . EBPs related to appropriate management of the third The largest burden of both overall neonatal, and, more stage of labor and neonatal care were observed in inter- specifically, preterm mortality, occurs within the first vention clinics receiving PRONTO training as compared 24 h of life . Similarly, a large proportion of stillbirths to controls in Mexico and Guatemala [13, 14]. are intrapartum deaths, occurring less than 12 h before Third, quality improvement (QI) is a strategy to delivery and thus resulting in infants without any signs optimize processes through testing of iterative changes. of maceration or skin deterioration (i.e., fresh stillbirths) QI methods to improve quality of care have been de- . Thus, the intrapartum and immediate postnatal pe- scribed with various frameworks, such as the riods represent critical windows of opportunity to im- Plan-Do-Study-Act cycles . In resource-limited set- prove neonatal outcomes in these settings. Estimates tings, QI approaches have been used to improve effective suggest that improved facility-based care during labor scale-up of EBPs. For example, the Project Fives Alive! is and birth and immediate newborn care can avert 0.8 a country-wide QI project in Ghana focused on the im- million newborn deaths by 2025 . These estimates re- proved delivery of maternal and child health and nutri- flect the potential of packages of interventions, rather tional interventions . This project brings together QI than a single intervention, to make significant improve- teams at each facility that are responsible for the devel- ments in outcomes. However, many proven interven- opment and testing of “change ideas.” Members of facil- tions are not widely used in many low- and ity teams form an Improvement Collaborative Network, middle-income countries (LMICs). Health system bottle- giving the opportunity for sharing of failures, successes necks, like financial resources and workforce capacity and ideas. Data from this study indicate that change ac- limitations, have constrained systems’ abilities to deliver tivities were associated with increased postnatal attend- some interventions at scale . ance among underweight infants . However, Several approaches have been shown to improve the systematic reviews have demonstrated that QI studies uptake of obstetric and neonatal evidence-based prac- have modest evidence and that more studies are needed tices (EBPs) and interventions. First, the WHO devel- to understand the necessary and sufficient elements of oped the Safe Childbirth Checklist which includes 29 QI strategies . EBPs that focus on maternal and neonatal outcomes at Lastly, in order to understand the impact of any inter- four pause points – on admission to the facility, at the ventions on maternal and neonatal outcomes, robust time of pushing (or before cesarean delivery), soon after measurement of data must be prioritized. The Every birth (within 1 h) and at discharge. This checklist has Newborn Action Plan (ENAP) published 10 core indica- been used in a variety of LMICs contexts, including tors that should be tracked in order to improve quality India, Sri Lanka, and Namibia [9–11]. In Namibia, the of care for mothers and newborns . This ENAP Otieno et al. Trials (2018) 19:313 Page 3 of 12 roadmap underscores the importance of standardizing Methods indicator definitions and strengthening routine health Trial design information systems. In the field of prematurity, in par- This study is a pair-matched, cluster randomized con- ticular, data strengthening around accurate GA using trolled trial (CRCT) among 20 public sector health facil- available resources and systems has been identified as a ities in the Busoga Region of Uganda (four facilities) and key issue. For example, in a study examining the use of in Migori County, Kenya (16 facilities, including 14 pub- antenatal corticosteroids in LMICs, birthweight of lic facilities and two not-for-profit missionary hospitals). below the 5th percentile was used as a proxy measure The full intervention package (data strengthening (DS), for preterm birth because GA recording was considered modified Safe Childbirth Checklist (mSCC), PRONTO of insufficient accuracy . provider training, and quality improvement (QI)) will be To our knowledge, no published studies have exam- introduced to 10 facilities (intervention arm); the ined the impact of a facility-based intervention package remaining 10 facilities will receive DS + mSCC (control focused on improving uptake of obstetric and neonatal arm) (Fig. 1). All facilities will begin with DS + mSCC EBPs in order to address preterm mortality. The East intervention components in order to capture preliminary Africa Preterm Birth Initiative (PTBi-EA) hypothesizes data for baseline and facility matching, as well as to that a facility-based intervention package administered standardize definitions of key indicators related to GA during the intrapartum and immediate postnatal period and newborn outcomes. Roll-out of mSCC and support will decrease the neonatal mortality rate among preterm will differ between the control and intervention sites, in neonates. Our package comprises four components: (1) that the latter will receive additional mSCC mentorship strengthening of routine data collection and data use ac- and support through synergies with PRONTO and QI. tivities, including regular data quality assessments The intervention package will be delivered at the facility (DQAs); (2) implementation of the WHO Safe Child- level, while outcomes will be measured at both an indi- birth Checklist modified for preterm birth; (3) PRONTO vidual and facility level. provider training and mentoring on key EBPs to In addition to the 10 pairs of matched facilities, three strengthen intrapartum and immediate newborn care; referral hospitals to which the respective sub-county or and (4) support of QI teams. We believe that these inter- district hospitals send their high-risk deliveries will re- ventions, in combination, will improve awareness and ceive the full intervention package. While these three re- practice of EBPs; teamwork, communication, and re- ferral hospitals are not included in the randomization spectful maternity care; and use of data for scheme, cases referred in from any one of the 20 facil- decision-making. We anticipate that results from this ities will be included in the primary outcome analysis. study will inform how interventions used in combination can improve personnel and facility capability and readi- Setting and site selection ness to respond to preterm labor and delivery, as well as The study regions of Migori County, Kenya and Busoga care for the preterm infant. Region, Uganda were selected based on in-country Fig. 1 Schematic of the study design Otieno et al. Trials (2018) 19:313 Page 4 of 12 stakeholder input. Prematurity burden and presence of reinforcing intervention components, PRONTO and QI. synergistic parallel maternal/newborn research or imple- Healthcare providers in intervention facilities who pro- mentation studies were considered. Health facilities were vide consent for PRONTO simulation trainings will also asked to participate in this study by in-country partners. be enrolled as study participants to ascertain changes in Formal approval from facility leadership was obtained knowledge and practices. Facility staff in the intervention before any activities commenced. Given that facilities facilities will be organized into QI teams to develop and were not selected from a target population of hospitals, implement QI programs. the intervention effects should be interpreted as impact evaluation of the intervention package implemented at Intervention package components the said facilities. A complete list of all facilities can be Ten intervention sites and the three referral-level hospi- found at the clinical trial registration (ClinicalTrials.gov, tals will receive an intervention package comprising DS ID: NCT03112018). + mSCC + PRONTO + QI, while the remaining 10 con- The Busoga Region of Eastern Uganda contains ap- trol facilities will receive DS + mSCC. Each intervention proximately three million people, or 10% of Uganda’s component is described in detail below. The interven- population, with over 80% of residents living on less tion package is designed to strengthen and reinforce than US$1 per day . The estimated preterm birth EBPs, as well as improve teamwork, communication, re- rate for Uganda is 13.57% and the neonatal mortality spectful maternity care and data use. All study activities rate is 21 per 1000 live births [3, 4]. Our selected six consist of known interventions or strategies. There are health facilities include approximately 22,000 deliveries no experimental interventions that would directly im- per year, with 9000 deliveries from the four hospitals pact patient safety. pair-matched in this study. Migori County, located in southwestern Kenya, has a population of approximately Data strengthening (DS) 917,170, wherein 43% of the population lives below the Improvements in measurement and data use in the poverty line . The estimated preterm birth rate for study sites are critical to establishing baseline measures Kenya is 12.30% and the neonatal mortality rate is 23 and achieving and demonstrating reductions in the bur- per 1000 live births [3, 4]. Our 17 selected Migori den of preterm birth. Therefore, we will begin our study County health facilities include approximately 10,000 de- by strengthening existing data collection processes in liveries per year, with 7500 deliveries at pair-matched health facilities, introducing standard tools to improve sites. GA assessment, and reviewing standardization of indica- tors based on national guidelines. We will also work Study population with in-country stakeholders to develop and iterate a In all facilities, women accessing delivery care services Data Dashboard to improve data use and dissemination who are admitted for labor will be eligible for this study. of our study data (Table 1). Anonymized data on all deliveries will be extracted from DS initial training will include review of these compo- maternity registers. For follow-up, mothers of newborns nents with facility clinical leadership, health records offi- who are discharged alive and born less than 2500 g or cers and district staff, followed by site trainings with between 2501 g and 3000 g while also being identified as maternity nurses and health record staff. Refresher DS less than 37 weeks by recorded GA in the maternity trainings will be offered as needed during the course of register are being asked to participate in this study and the study. Intermittent DQAs will be implemented every approached for consent for follow-up for 28-day out- 6–12 months to collect data on specific DHIS-related in- come. These inclusion parameters were selected based dicators to assess gaps between reported and actual indi- on baseline data showing a poor correlation between cators (e.g., errors in transcription) across all control birthweight and reported GA in the maternity register. and intervention facilities. This process will also help Women or newborns from enrolled sites who are re- identify barriers in the reporting processes and flow. ferred to one of the three referral facilities will remain in the study. Their outcomes will be allocated to the facility Modified Safe Childbirth Checklist to which the woman first presented. Each country team will modify the WHO Safe Childbirth Healthcare workers providing labor and delivery and Checklist in order to adhere to their national guidelines. immediate newborn care services at referral, control, It will be adapted to the local setting and modified to and intervention facilities will participate in DS initial emphasize identification of preterm labor and care of and refresher workshops, DQAs, as well as initial in- the preterm infant. Specifically, we will incorporate add- struction and minimal refreshers or reorientation on use itional elements focused on GA assessment and docu- of the mSCC. Intervention sites will have ongoing sup- mentation, prematurity-related intrapartum/immediate port for mSCC utilization, in addition to the other postnatal care practices (e.g., use of magnesium sulfate, Otieno et al. Trials (2018) 19:313 Page 5 of 12 Table 1 Components of data strengthening Data strengthening component Description Reinforcing current status of data systems and indicators We will provide technical support to standardize definitions of indicators currently being collected, improve adherence to national guidelines on labor and delivery documentation of registers, improve quality of reporting, and strengthen existing data quality assurance and data use processes. Follow-up assessments to gauge improvements in facility systems including data quality assessments (DQAs) will be conducted at regular intervals. Refining standardized gestational age measurement We will strengthen the use of last menstrual period (LMP) with pregnancy wheels to accurately calculate gestational age (GA). We will reinforce more accurate measures of birthweight by providing training and assessing calibration of facility scales at regular intervals. Developing a Data Dashboard To improve data use and dissemination of routine data, we will create a synchronized online Data Dashboard repository system that is adaptable for local providers, health officials, and national policymakers. This tool will be based on discussions among various PTBi stakeholders to better understand and respond to data needs, and is also integrated in QI and project monitoring and evaluation. antenatal corticosteroids, immediate skin to skin, etc.). Selection of PRONTO mentors/trainers will be con- We will also include an additional pause point at initial ducted in each country, with an initial pool of up to 15 presentation or triage (i.e., before a woman is admitted candidates from which we will select the 5–10 highest for labor), as well as prompts focused on ascertaining performing trainers. Due to the overlapping clinical and additional maternal demographic information, clinical curricular content between the two countries, refreshers risk factors and history. will be conducted as joint facilitator training for the Each country’s mSCC will be piloted in order to Kenyan mentors and Ugandan trainers. However, the optimize content and roll-out. The mSCC will be intro- training mode of delivery will vary between Kenya and duced during initial DS activities at all study sites. It is Uganda. Kenya will utilize an in situ mentoring program intended to serve as both a decision aid for providers of whereby each intervention facility will receive key EBPs, as well as a data source for the study. An high-intensity/4-day per week mentorship and a pair of mSCC will be included in the maternity inpatient record mentors will rotate among intervention sites during the for each woman in all control and intervention facilities. study duration. They will spend a combined total of 9– After piloting, study data staff will review all maternity 12 weeks at each intervention facility over the study dur- charts of cases eligible for follow-up each month and ab- ation and visits will include bedside mentoring, stract a selected number of essential data variables. video-recorded, in situ simulations, knowledge reviews, Study personnel will also monitor mSCC completeness skills stations, teamwork activities, and mSCC support. and uptake by each of the five pause points either by In Uganda, a high-intensity/shorter modular strategy will convenience or purposive sampling. These data will be be used. A modular-based training program will be displayed on the Data Dashboard quarterly, and will paired with 2-day-long in-situ simulation refresher and allow the study teams to tailor the mSCC approach de- training visits. Modules and refresher/training visits will pending on uptake and use. be spread out during the study period, and will similarly amount to approximately 6 weeks of mentorship. Thus, PRONTO simulation-based training while the mode of delivery for training in each country The Kenya-Uganda unified PRONTO emergency obstet- will be different, provider teams in each country will re- ric and neonatal care simulation training will emphasize ceive approximately 56–58 h of PRONTO-based instruc- the identification, triage, and management of preterm tion using the same curriculum. labor and birth with a curriculum specifically adapted for this context. It will include strengthening preterm labor identification with more accurate GA assessments, Quality improvement (QI) cycles intrapartum care, and immediate management of fragile Each facility in the intervention arm will have a desig- infants. The training also emphasizes identification and nated QI team comprising facility leadership, nurses, management of preeclampsia, chorioamnionitis, and and health record staff (five to seven people). If teams other conditions related to preterm birth. The mSCC have been trained previously through other QI efforts, will be integrated into all PRONTO clinical activities to we will revitalize and support these ongoing efforts in provide facility staff with continued opportunities to intervention sites. Otherwise, we will offer foundational reinforce its use. Any change ideas that arise from these training in QI methods. These teams will carry out PRONTO activities will be integrated into QI efforts. Plan-Do-Study-Act cycles which include identification of Otieno et al. Trials (2018) 19:313 Page 6 of 12 a problem or bottleneck in the facility, implementation allocation will be done using a computerized of solutions, tracking of the outcomes of the changes, random number generator by the statistical team and implementation or adjustment based on the results. with no contact or direct interest in any specific QI teams across facilities (known collectively as the QI facilities. As this is a cluster trial, there will be no collaborative) will also participate in a learning session blinding or allocation concealment. every 3 to 6 months to discuss core learnings and QI in- Step 3: Upon accruing sample size, the control arm dicators. QI indicators will be chosen by each country facilities will be assigned to the full intervention, and will focus on EBPs expected to result in decreased adding PRONTO training and QI cycles. neonatal mortality among preterm infants, such as up- take of Kangaroo Mother Care, antenatal corticosteroid Figure 2 depicts the schedule of enrollment, interven- provision, and breastfeeding. At these learning sessions, tions, and assessments to be conducted. QI teams across facilities will be able to share progress on QI indicators, lessons learned, and best practices. Outcomes Elements of the package, namely PRONTO, the Data The primary outcome for this study is 28-day mortality Dashboard and the mSCC, will be integrated with QI ef- rate among preterm infants. For this primary analysis, forts. First, the Data Dashboard will help generate spe- the denominator will include: cific visual data to inform facility teams on progress and remaining performance gaps with respect to QI indica- All live births weighing greater than 1000 g and less tors and selected EBPs. Second, areas of possible im- than 2500 g provement that arise through PRONTO mentorship and All live births weighing between 2501 g and 3000 g simulation will be shared with QI teams. Lastly, the with a documented GA of less than 37 weeks mSCC may serve as a data source to document QI All fresh stillbirths weighing greater than 1000 g and indicators. less than 2500 g All fresh stillbirths weighing between 2501 g and 3000 g with a documented GA of less than 37 weeks Intervention roll-out The approach for roll-out of the intervention package This outcome will be measured by comparing mortality will be as follows: rates of preterm infants at 28 days after delivery across the intervention and control arms to determine the effect of Baseline: Across all selected facilities, preliminary DS the package of facility-level interventions. The upper limit with initial introduction of the mSCC will allow us of 3000 g (which is coupled to a documented GA of less to conduct a baseline survey to measure incidence of than 37 weeks) was agreed upon based on preterm births and preterm mortality in all facilities INTERGROWTH-21st standards. At this weight, we as- (at least 2 months). Baseline data will also be sume that we will capture 90% and 97% of < 34-week in- collected on key variables that are predictive/ fants and 60% and 70% of < 37-week infants, for female correlated with the primary outcome of mortality and male children, respectively . The lower limit of in- rate among preterm infants, including but not clusion for the primary analysis was set at 1000 g to align limited to number of deliveries, number of low with the International Classification of Diseases 10th revi- birthweight babies, number of neonatal deaths, sion (ICD 10) definition of late fetal death (i.e., birthweight number of stillbirths, presence of a newborn corner, of > 1000 g or > 28 weeks’ gestation) . However, data on personnel capacity for intrapartum/postnatal care, infants with signs of life born weighing < 1000 g will be etc. followed for secondary data analyses, including outcomes Step 1: Facilities will be pair-matched using key fac- at 28 days of life. Selected secondary outcomes are listed in tors correlated with outcomes. Specifically, sites will Table 2.We include additional information as to how they be pair-matched based on variables collected during will be measured and how often during the study duration. baseline that are predictive or correlated with the primary outcome of mortality rate among preterm neonates (i.e., delivery volume, number, and type of Data collection/quality control providers, facility-based neonatal mortality rate). For the primary outcome, existing facility-based registers Step 2: Within each matched pair, one facility will be will be used as primary data sources. Data entry into regis- randomized to the intervention arm and the other tries is conducted by facility care providers, as part of will be randomized to the control arm. Intervention existing routine data collection. Information in these facilities will receive PRONTO and QI interventions registries will then be extracted by study personnel. Study in addition to DS and mSCC. Randomization and staff will visit each facility at least once per month to Otieno et al. Trials (2018) 19:313 Page 7 of 12 Fig. 2 Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) Figure collect routine facility data from register reviews. All data Data strengthening will be uploaded via Open Data Kit via tablet or laptop. Study personnel in each country will collect data from For preterm babies discharged alive from hospital, maternity registers on a monthly basis. In addition to mothers will be consented to be followed up by phone reviewing the quality of records, they will also generate at 28 days following delivery. Consent of eligible summary reports for data sharing among the research mothers and their newborns will take place prior to dis- team and facility leads. charge or referral. Contact information will be derived Routine indicators and process indicators for QI cycles from consent forms, and from the mSCC as needed. will be collected and displayed on a Data Dashboard ac- Outcomes will be determined by targeted follow-up of cessible to study staff, intervention and referral health study participants via phone call. Where phone calls are facility staff, and county health authorities. While these insufficient to trace mothers, the Kenya team will engage Data Dashboard displays will be customized to different with community health volunteers and Uganda will em- audiences, all of the data included will be aggregate, and ploy community outreach nurses to seek out mothers. no individual data will be displayed. Otieno et al. Trials (2018) 19:313 Page 8 of 12 Table 2 Select secondary outcomes Secondary outcome How variable will be measured Time frame Data quality of key indicators in facility-based GA-birthweight concurrence, DQAs and mSCC Baseline through study completion, an registers (includes GA, facility discharge status, data review or QI indicators average of 18 months; at least quarterly preterm birth incidence) Pre-discharge mortality among all infants > Clinical record review at facility Baseline and through study completion, an 1000 g average of 18 months Facility-based maternal mortality Clinical record review at facility Baseline and through study completion, an average of 18 months Perinatal mortality (fresh stillbirths and deaths Parental report at 28 days and clinical record Baseline and through study completion, an within 7 days) among eligible preterm infants review at facility average of 18 months (includes pre-discharge mortality) Mortality among preterm infants and those Parental report at 28 days and clinical record Baseline and through study completion, an born alive between 500 g and 999 g at birth review at facility at time of first contact average of 18 months (include pre-discharge mortality and 28-day mortality) Average number of EBPs or Ministry of Health Measured in PRONTO simulation videos, Baseline and through study completion, an management guidelines demonstrated in observed live births and/or mSCC. To be average of 18 months. PRONTO administered simulated case videos and live birth complemented by pre-post knowledge tests at pre-determined time points observations and facility assessments Knowledge improvement of EBPs PRONTO pre-post knowledge test scores PRONTO administered at pre-determined time points Proportion of eligible cases receiving EBPs QI indicators reported at learning sessions Baseline and through study completion, an reported by QI teams (QI indicators include average of 18 months; QI learning sessions Kangaroo Care, antenatal corticosteroid held every 3–6 months provision, and breastfeeding) Facility readiness to handle delivery and Measured by a facility assessment tool Bi-annually over the study period newborn complications Understanding of contextual factors influence Process evaluation that incorporates To be conducted mid-study implementation of strengthening maternal document review, surveys and qualitative and newborn care interventions interviews and focus groups Prevalence of preterm birth phenotypes in the Retrospective and/or prospective chart review Baseline and through study completion; at study sites least annually in select sites GA gestational age, DQA data quality assessment, EBP evidence-based practice, mSCC Modified Safe Childbirth Checklist, QI quality improvement Modified Safe Childbirth Checklist (mSCC) PRONTO’s on-site simulation training program, we will The mSCC will be distributed to facilities and fixed into collect video-recordings of simulated birth scenarios and the patient charts in readiness for use by the healthcare debriefs conducted in participating hospitals led by providers. Staff will be adequately trained on the use of PRONTO-trained mentors/trainers. These videos will be the checklist and regular reinforcement conducted as coded using Studiocode™ software to create scores based scheduled at intervention sites. Study staff will review all on how often EBPs are practiced in simulation and if maternity charts for newborns who meet our eligibility this changes over time. criteria. Each month, they will abstract key data from eli- gible admissions from the mSCC in order to compile Quality improvement (QI) cycles coverage indicators for key interventions and EBPs. Up- We will track process indicators of these QI cycles, such as take and completeness of the mSCC will also be number of projects started, number of goals reached, and determined. amount of change detected by the QI team in studying their implementation. We will implement a QI documentation PRONTO simulation-based training journal for the sites. QI teams will also track key EBPs, such To measure changes in knowledge through training and as Kangaroo Care uptake, and track their progress against it. mentoring, we will conduct evaluations in the form of pre- and post-knowledge tests of PRONTO-trained mid- Data management wives, nurses, and physicians before and after each train- Data from registries will be collected using a secure ing session and periodically during mentoring visits. database via the Open Data Kit data entry platform and These evaluations will be adapted to the local context hosted on a secure server. Data will be reviewed for ac- based on previously developed knowledge assessment curacy and completeness by a data manager before tools used by PRONTO. To evaluate the impact of entry, and the data entry system will include automated Otieno et al. Trials (2018) 19:313 Page 9 of 12 range and logic checks to identify any data entry mis- newborn) in order to improve precision. We will directly takes before they are saved. All devices used for data incorporate knowledge of the pair-matched entry (laptops or tablets) will be encrypted and password randomization scheme into estimation by making the protected. targeting stage of this procedure a function of this The research team and stakeholders will have access known assignment mechanism. Primary outcome data to aggregate data across facilities through the Data analysis will be conducted in collaboration among Dashboard. For example, each facility will have access to in-country partners and UCSF. their own data including 28-day outcome, but stratified Secondary analyses will also be performed, in some data by control and intervention data will not be shared. cases contrasting secondary outcomes between the inter- Moreover, only the study biostatistician and core team vention and control groups, and in others contrasting will have access to the unblinded dataset prior to study these outcomes pre and post intervention within the completion. intervention groups. Secondary analyses will primarily be descriptive, comparing means or proportions between Sample size and power calculations groups, trends over time, composite scores as appropri- Our primary analysis will combine data across our se- ate for each measure. Both primary and secondary ana- lected facilities in Kenya and Uganda. Since we will ex- lyses will be conducted in R or Stata. Process data clude the referral hospitals from our primary analysis, including qualitative interviews or reports will be ana- the project takes place in 20 facilities with an expected lyzed by hand or in Atlas ti, identifying and grouping volume of 46,000 deliveries over 24 months. In an initial themes that emerge. calculation, prior to baseline data collection, we assumed Additional analyses will be conducted for each inter- an average preterm birth rate of 12% and expected to vention component. For example, PRONTO-related see about 5500 preterm deliveries within this period. We knowledge will be assessed against the standard guide- assumed a 25% loss-to-follow-up rate for eligible cases. lines for management. Simulation and debrief videos will This yields at least 200 projected preterm deliveries per be analyzed using Studiocode™ software which enables facility with known 28-day mortality outcome. the systematic coding of videos to measure of changes in Detectable effect sizes were estimated by standard t use of EBPs, as well as teamwork and communication test procedures adjusted to account for the design effect techniques. Video analysis of simulation and debrief re- due to clustering of outcomes within facilities to attain sults will be shared with participants in the form of 80% power while maintaining type I error at 5%. At a structured feedback to PRONTO mentors on their 25% 2-year cumulative incidence of 28-day mortality simulation and debriefing facilitation skills. across both countries in the absence of the intervention, this would allow us to detect a 25% reduction in cumu- Result dissemination lative incidence if the between-cluster outcome coeffi- In accordance with the Bill & Melinda Gates Foundation cient of variation is 0.2 or below. If this coefficient open-access policy, we will publish in open-access jour- increases to 0.3, we would be powered to detect a 30% nals. The final trial dataset will be made publicly available reduction. If it increases to 0.4, we would be powered to after study completion once all datasets are cleaned and detect a 40% reduction. initial results are reported. We plan to disseminate evalu- ation findings to both internal and external stakeholders, Analyses plans including facility staff implementers, Ministries of Health For the primary outcome, the analysis will contrast mor- policymakers, and the Bill & Melinda Gates Foundation. tality at 28 days among preterm infants between the intervention and control groups. This will be performed Ethical considerations using hierarchical, targeted maximum-likelihood estima- The intervention package is designed to strengthen and tion which accounts for within-cluster correlation by reinforce best EBPs, and all study activities consist of controlling for cluster-level covariates . The baseline known interventions or strategies. There are no experimen- covariates we will measure for each facility include deliv- tal interventions that would directly impact patient safety. ery volume, baseline neonatal mortality rate among pre- Should any adverse events be reported, these will be imme- term infants, preterm birth rate, and country. diately reported to study leadership and ethics committees. Additionally, targeted maximum-likelihood estimation will allow us to incorporate individual-level baseline co- Approvals variate information (such as date of presentation, mater- This proposal was submitted to KEMRI, Makerere Uni- nal age, parity/gravidity, HIV status, maternal and fetal versity School of Public Health, and UCSF Scientific and complications at presentation, last menstrual Ethical Research Bodies for scientific and ethical ap- period (LMP), infant birthweight, final diagnosis of proval before study initiation. Otieno et al. Trials (2018) 19:313 Page 10 of 12 Consent procedures Discussion This is an implementation science study and intervention This study focuses on facility-based care of mothers and components will be applied to facilities rather than individ- infants during the intrapartum and immediate newborn uals. In most cases, there will be no direct contact with the period. Across Kenya and Uganda, as well as each of the participants except for the 28-day follow-up. For the pri- study facilities, the intervention components will be tai- mary outcome, risks to participants will be minimized by lored by in-country stakeholders. For example, the the fact that facility registers and medical records will be mSCC and PRONTO curriculum will be adapted to en- used as the primary source of information and no identifi- sure adherence to national guidelines. If ongoing QI is able information will be collected or used. Data abstraction already in place at our selected intervention facilities, QI from registers, medical records and the mSCC will be con- teams and strategies will synergize with these existing ef- ducted in a private, confidential area of each facility. How- forts or program. This will ensure alignment of the trial ever, for 28-day follow-up among eligible newborns, to current policies and practices. mothers will be asked to provide written consent prior to The trial, as designed, will assess a package of inter- discharge. No incentives will be provided, and women can ventions rather than a single intervention. This approach opt out from participation. reflects a consensus in the field that multiple interven- For providers in the intervention arm, both PRONTO tions are needed, but creates challenges for researchers mentors and mentees will be asked for written consent accustomed to being able to attribute effects more gran- authorizing the use of knowledge tests and video data for ularly to specific interventions. We believe that each analysis. While QI indicators and a change in direction of intervention component reinforces each other to create performance would be noted over the course of the study, a synergistic package that, when implemented together, QI team members and facility staff will not be asked for will allow facility providers to improve EBPs and their consent as no identifying individual data will be collected. appropriate documentation (Fig. 3). The interventions Fig. 3 Logic model for study interventions Otieno et al. Trials (2018) 19:313 Page 11 of 12 target different types of facility staff and cadres of pro- Trial status viders, demonstrating that shared and cooperative strat- This trial has completed planning and began enrollment egies are needed to address quality of care. in October 2016. Study enrollment is ongoing with ex- pected completion by October 2018. Challenges and opportunities Several challenges will need to be overcome in order to Trial registration ensure the success of this study. First, GA accuracy re- Preterm Birth Initiative Kenya/Uganda Protocol dated: 19 mains to be a major obstacle in these settings. Lack of December 2016; ClinicalTrials.gov, ID: NCT03112018. early ultrasound, coupled with late antenatal care-seeking practices and inherent challenges with using LMP, makes Additional file GA difficult to robustly capture. Second, using routine data sources may uncover challenges with data quality Additional file 1: Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) Checklist. (DOC 123 kb) and completeness. While data strengthening is a founda- tional intervention across all of our sites, staff turnover, in- Abbreviations dustrial strikes, burnout, and other contextual factors may DQA: Data quality assessment; DS: Data strengthening; EBP: Evidence-based impact data. Third, although this is a package of interven- practice; ENAP: Every Newborn Action Plan; GA: Gestational age; LMIC: Low- tions, each of the component parts has its unique and middle-income country(ies); LMP: Last menstrual period; mSCC: Modified Safe Childbirth Checklist; PTBi-EA: East Africa Preterm Birth Initiative; strengths and challenges. Uptake and acceptability of the QI: Quality improvement; WHO: World Health Organization individual components may vary. Lastly, our trial focuses on improving facility-based care during the intrapartum Acknowledgements With gratitude, the authors acknowledge the contributions of the PTBi-UCSF, and immediate newborn period; however, our primary PTBi-Kenya, and PTBi-Uganda study teams. We thank Leah Kirumbi, Nelly outcome is 28-day mortality among preterm infants (de- Mugo, Anthony Wanyoro, Darious Kajjo, Roger Myrick, Rikita Merai, Lara Miller, nominator defined as the sum of live born infants weigh- Wenjing Zheng, and Alejandra Benitez. We are grateful for the collaboration of the Ministries of Health of Kenya and Uganda, as well as the local government ing greater than 1000 g and less than 2500 g or GA of less representatives of Migori County, Kenya and the Busoga Region, Uganda for than 37 weeks and weight less than 3000 g plus fresh still- their support to conduct this trial in selected health facilities. We thank the PTBi births with the same weight and GA criteria). The ability East Africa External Advisory Committee members who have provided insight and suggestions along the way: Pierre Barker, Zulfiqar Bhutta, Colin Boyle, Alex to monitor post-discharge health or activities will not be Coutinho, Eric Goosby, Linda Guidice, Jerker Liljestrand, Suellen Miller, Jaime possible within the scope of this project. Sepulveda, and Marleen Temmerman. We also extend our gratitude to key Several research opportunities arise within this consulting partners who have shaped and deliver the key interventions: Nana Twum-Danso, building on her work with IHI and Project Fives Alive and the trial. First, as preterm birth is often described as a PRONTO International team. We are also indebted to the WHO and the Better syndrome , this study will allow for nested stud- Birth Initiative for the groundwork done on the Safe Childbirth Checklist. ies examining maternal, fetal, and placental risk fac- Funding tors that contribute to preterm birth. As such, This trial is supported by the East Africa Preterm Birth Initiative, a multi-year, preterm birth phenotyping studies using the frame- multi-country effort generously funded by the Bill & Melinda Gates Foundation. work described by Barros et al. (2012) will be nested The funders reviewed the study design and will not have input on study analysis or interpretation. at some of our sites . Second, as GA assessment will prove to be a challenge in many of our sites, this Availability of data and materials poses a unique opportunity to test other measures to Final dataset related to this study protocol will be made available at a minimum of UCSF’s Datashare platform in accordance with the Open Access more accurately estimate GA. This may include, for terms of our funding. example, studies focused on discovery of biomarkers and/or comparison of various anthropometric Authors’ contributions PO, PW, and DW serve as co-PIs of this study and conceived of the study measures. and its design. EB, GN, and KA serve as study program managers and oversee This study describes a single CRCT that spans regions trial implementation with PO, PW, and DW. GN, KA, EB, and NS contributed in Western Kenya and Eastern Uganda. The interven- to the development of the protocols and study tools; FL participated in the development of the mSSC tools; RK contributed to study tools and procedures tions implemented will need to be tailored and adapted related to data collection, abstraction, and follow-up. NS and EB participated in to the local context and national guidelines. It will also the drafting of this manuscript. All authors read and approved the final be important to document any overlapping or contribut- manuscript. ing factors of other ongoing or newly introduced mater- Ethics approval and consent to participate nal or newborn health initiatives that may impact our The study protocol (version 1.0) has been approved by the University of study results, such as other QI initiatives or training pro- California, San Francisco Institutional Review Board (Study no: 16–19,162), grams. Nonetheless, this is a great opportunity to dem- Kenyan Medical Institute Scientific and Ethics Review Unit (SERU protocol no: KEMRI/SERU/CCR/0034/3251), and the Makerere University Higher Degrees, onstrate both the feasibility and challenges associated Research, and Ethics Committee (Protocol ID: IRB00011353). The protocol will with adapting interventions under a shared research be reviewed and reapproved on an annual basis and if protocol amendments study and outcome measure. are needed. Additional file 1 describes the SPIRIT recommended items to Otieno et al. Trials (2018) 19:313 Page 12 of 12 address in a clinical trial protocol. The trial is registered at ClinicalTrials.gov, Program in India. N Engl J Med. 2017;377(24):2313–24. https://doi.org/10. ID: NCT03112018. Consent procedures are described above in “Ethical 1056/NEJMoa1701075. considerations”; for 28-day follow-up among eligible newborns, mothers will be 13. Walton A, Kestler E, Dettinger J, Zelek S, Holme F, Walker D. Impact of a asked to provide written consent prior to discharge. low-technology simulation-based obstetric and newborn care training scheme on non-emergency delivery practices in Guatemala. Int J Gynecol Competing interests Obstet. 2016;132:359–64. PO and PW have no financial or non-financial competing interests to declare. 14. Fritz J, Walker DM, Cohen S, Angeles G, Lamadrid-Figueroa H. Can a DW is co-founder of PRONTO International. simulation-based training program impact the use of evidence based routine practices at birth? Results of a hospital-based cluster randomized trial in Mexico. PLoS One. 2017;12(3):e0172623. Publisher’sNote 15. Taylor MJ, McNicholas C, Nicolay C, Darzi A, Bell D, Reed J. Systematic Springer Nature remains neutral with regard to jurisdictional claims in review of the application of the plan-do-study-act method to improve published maps and institutional affiliations. quality in healthcare. BMJ Qual Saf. 2014;23:290–8. 16. Twun-Danso N, Akanlu G, Osafo E, Sodzi-Tettey S, Boadu RO, Atinbire S, et Author details al. A nationwide quality improvement project to accelerate Ghana’s 1 2 Kenya Medical Research Institute, Nairobi, Kenya. Department of Health progress towards Millennium Development Goal 4: design and Policy, Planning and Management, Makerere University School of Public implementation progress. Int J Qual Health Care. 2012;24(6):601–11. https:// Health, Kampala, Uganda. Center of Excellence for Maternal, Newborn and doi.org/10.1093/intqhc/mzs060 Child Health, Makerere University, Kampala, Uganda. Institute for Global 17. Singh K, Speizer I, Handa S, Boadu R, Atinbire S, Barker P, Twum-Danso N. Health Sciences, University of California, San Francisco, CA, USA. Department Impact evaluation of a quality improvement intervention on maternal and of Obstetrics, Gynecology, and Reproductive Sciences, University of California, child health outcomes in Northern Ghana: early assessment of a national San Francisco, CA, USA. scale-up project. Int J Qual Health Care. 2013;25(Issue 5):477–87. https://doi. org/10.1093/intqhc/mzt054. Received: 5 February 2018 Accepted: 17 May 2018 18. Schouten L, Hulscher M, Everdingen J, Huijsman R, Grol R. Evidence for the impact of quality improvement collaboratives: systematic review. BMJ. 2008; 336:1491. References 19. Moxon S, Ruysen H, Kerber K, Amouzou A, Fournier S, Grove J, et al. Count 1. Liu L, Oza S, Hogan D, Perin I, Rudan I, Lawn JE, Cousens S, Mathers C, Black every newborn; a measurement improvement roadmap for coverage data. RE. Global, regional, and national causes of child mortality in 2000–13, with BMC Pregnancy Childbirth. 2015;15(Suppl 2):S8. projections to inform post-2015 priorities: an updated systematic analysis. 20. Berrueta M, Hemingway-Foday J, Thorsten V, Goldenberg R, Carlo W, Garces Lancet. 2015;385(9966):430–40. https://doi.org/10.1016/S0140- A, et al. Use of antenatal corticosteroids at health facilities and communities 6736(14)61698-6. Epub 2014 Sep 30 in low-and-middle income countries. Reprod Health. 2016;13:66. 2. Blencowe H, Lee A, Cousens S, Bahalim A, Narwal R, Zhong N, et al. Preterm 21. People living below national poverty line. World Bank. http://povertydata. birth-associated neurodevelopmental impairment estimates at regional and worldbank.org/poverty/country/KEN; http://povertydata.worldbank.org/ global levels for 2010. Pediatr Res. 2013;74(Suppl 1):17–34. https://doi.org/ poverty/country/UGA. 10.1038/pr.2013.204. 22. Villar J, Ismail L, Victoria CG, Ohuma E, Bertino E, Altman D, et al. 3. Mortality rate, neonatal (per 1000 live births). 2015. Accessed on 20 Jul 2015 International standards for newborn weight, length, and head from http://data.worldbank.org/indicator/SH.DYN.NMRT. circumference by gestational age and sex: the Newborn Cross-Sectional 4. March of Dimes, PMNCH, Save the Children, WHO. In: Howson CP, Kinney Study of the INTERGROWTH-21st Project. Lancet. 2014(384):857–68. MV, Lawn JE, editors. Born too soon: The Global Action Report on Preterm 23. Balzer L, Zheng W, van der Laan M, Pertersen M. A new approach to Birth. Geneva: World Health Organization; 2012. hierarchical data analysis: targeted maximum likelihood estimation of 5. Stillbirth rate (per 1000 total births). 2015. Retrieved from UNICEF: https:// cluster-based effects under interference. eprint arXiv:1706.02675. https:// data.unicef.org/wp-content/uploads/country_profiles/Kenya/country%20 arxiv.org/pdf/1706.02675.pdf. profile_KEN.pdfhttps://data.unicef.org/wp-content/uploads/country_profiles/ 24. Romero R, Dey S, Fisher S. Preterm labor: one syndrome, many causes. Uganda/country%20profile_UGA.pdf. Accessed 15 Nov 2017. Science. 2014;345:760–5. https://doi.org/10.1126/science.1251816. 6. Lawn JE, Blencowe H, Waiswa P, Amouzou A, Mathers C, Hogan D, et al. 25. Barros FC, Papageorghiou AT, Victora CG, Boble J, Pang R, Iams J, et al. The Stillbirths: rates, risk factors, and acceleration towards 2030. Lancet. 2016; distribution of clinical phenotypes of preterm birth syndrome: implications 387((10018)):587–603. for prevention. JAMA Pediatr. 2015;169(3):220–9. https://doi.org/10.1001/ 7. Bhutta Z, Das J, Bahl R, Lawn JE, Salam R, Paul V, et al. Can available jamapediatrics.2014.3040. interventions end preventable deaths in mothers, newborn babies, and stillbirths, and at what cost? Lancet. 2014;384(9940):347–70. https://doi.org/ 10.1016/S0140-6736(14)60792-3. 8. Dickson KE, Simen-Kapeu A, Kinney MV, Huicho L, Vesel L, Lackritz E, et al. Every newborn: health-systems bottlenecks and strategies to accelerate scale-up in countries. Lancet. 2014;384(9941):438–54. https://doi.org/10. 1016/S0140-6736(14)60582-1. Epub 2014 May 19 9. Semrau KEA, Hirschhorn LR, Kodkany B, Spector J, Tuller D, King G, et al. Effectiveness of the WHO Safe Childbirth Checklist program in reducing severe maternal, foetal, and newborn harm in Uttar Pradesh, India: study protocol for a matched-pair, cluster-randomized controlled trial. Trials. 2016; 17(1):576. 10. Patabendige M, Senanayake H. Implementation of the WHO safe childbirth checklist program at a tertiary care setting in Sri Lanka: a developing country experience. BMC Pregnancy Childbirth. 2015;15:12. https://doi.org/ 10.1186/s12884-015-0436-0. 11. Kabongo L, Gass J, Kivondo B, Nabihah K, Semrau KEA, Hirschhorn LR. Implementing the WHO Safe Childbirth Checklist: lessons learnt on a quality improvement initiative to improve mother and newborn care at Gobabis District Hospital, Namibia. BMJ Open Qual. 2017;6(2):e000145. https://doi. org/10.1136/bmjoq-2017-000145. eCollection 2017 12. Semrau KEA, Hirschhorn LR, Marx Delaney M, Singh VP, Saurastri R, Sharma N, et al. Outcomes of a coaching-based WHO Safe Childbirth Checklist
Trials – Springer Journals
Published: Jun 5, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera