UBC Faculty Research and Publications

MAGnesium sulphate for fetal neuroprotection to prevent Cerebral Palsy (MAG-CP)—implementation of a national… De Silva, Dane A; Synnes, Anne R; von Dadelszen, Peter; Lee, Tang; Bone, Jeffrey N; Magee, Laura A Jan 11, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13012_2017_Article_702.pdf [ 1.89MB ]
JSON: 52383-1.0366929.json
JSON-LD: 52383-1.0366929-ld.json
RDF/XML (Pretty): 52383-1.0366929-rdf.xml
RDF/JSON: 52383-1.0366929-rdf.json
Turtle: 52383-1.0366929-turtle.txt
N-Triples: 52383-1.0366929-rdf-ntriples.txt
Original Record: 52383-1.0366929-source.json
Full Text

Full Text

RESEARCH Open AccessMAGnesium sulphate for fetalneuroprotection to prevent Cerebral Palsy(MAG-CP)—implementation of a nationalguideline in CanadaDane A. De Silva1,2, Anne R. Synnes2,3, Peter von Dadelszen4,5, Tang Lee1,2, Jeffrey N. Bone1,2, MAG-CP, CPNand CNN collaborative groups and Laura A. Magee4,5*AbstractBackground: Evidence supports magnesium sulphate (MgSO4) for women at risk of imminent birth at < 32–34 weeksto reduce the likelihood of cerebral palsy in the child. MAGnesium sulphate for fetal neuroprotection to prevent CerebralPalsy (MAG-CP) was a multifaceted knowledge translation (KT) strategy for this practice.Methods: The KT strategy included national clinical practice guidelines, a national online e-learning module and, at MAG-CPsites, educational rounds, focus group discussions and surveys of barriers and facilitators. Participating sites contributed dataon pregnancies with threatened very preterm birth. In an interrupted time-series study design, MgSO4 use forfetal neuroprotection (NP) was tracked prior to (Aug 2005–May 2011) and during (Jun 2011–Sept 2015) the KTintervention. Effectiveness of the strategy was measured by optimal MgSO4 use (i.e. administration when andonly when indicated) over time, evaluated by a segmented generalised estimating equations logisticregression (p < 0.05 significant). Secondary outcomes included maternal effects and, using the Canadian NeonatalNetwork (CNN) database, national trends in MgSO4 use for fetal NP and associated neonatal resuscitation. With ananticipated recruitment of 3752 mothers over 4 years at Canadian Perinatal Network sites, we anticipated > 95% powerto detect an increase in optimal MgSO4 use for fetal NP from < 5 to 80% (2-sided, alpha 0.05) and at least 80% powerto detect any increases observed in maternal side effects from RCTs.Results: Seven thousand eight hundred eighty-eight women with imminent preterm birth were eligible for MgSO4 for fetalNP: 4745 pre-KT (18 centres) and 3143 during KT (11 centres). The KT intervention was associated with an 84% increase inthe odds of optimal use (OR 1.00 to 1.84, p< 0.001), a reduction in the odds of underuse (OR 1.00 to 0.47, p< 0.001) and anincrease in suboptimal use (too early or at ≥ 32 weeks; OR 1.18 to 2.18, p< 0.001) of MgSO4 for fetal NP. Maternal hypotensionwas uncommon (7/1512, 0.5%). Nationally, intensive neonatal resuscitation decreased (p=0.024) despite rising MgSO4 use forfetal NP (p< 0.001).(Continued on next page)* Correspondence: Laura.A.Magee@kcl.ac.ukAn abstract related to the content of this paper was accepted forpresentation at the Canadian National Perinatal Research Meeting,Montebello, CA (Feb 14–17, 2017), and the Royal College of Obstetriciansand Gynecologists Congress, Cape Town, ZA (Mar 20–22, 2017).4Department of Women and Children’s Health, St Thomas’ Hospital, 10thFloor, North Wing, Westminster Bridge Road, London SE1 7EH, UK5School of Life Course Sciences, Faculty of Life Sciences and Medicine, King’sCollege London, London, UKFull 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 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe 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.De Silva et al. Implementation Science  (2018) 13:8 DOI 10.1186/s13012-017-0702-9(Continued from previous page)Conclusion:Multifaceted KT was associated with significant increases in use of MgSO4 for fetal NP, with neither importantmaternal nor neonatal risks.Keywords: Preterm birth, Fetal neuroprotection, Magnesium sulphate, Cerebral palsy, Knowledge translation, Implementation,Interrupted time-seriesBackgroundComplicating approximately 10% of births, prematurityremains a major cause of perinatal mortality and mor-bidity, especially cerebral palsy (CP) [1–3]. Althoughsurvival rates of babies born preterm have risen, therehas been no parallel fall in neurodevelopmental impair-ment rates, especially among babies born very pretermat < 32 weeks’ gestation [4].By 2009, meta-analyses of randomised controlled trials(published between 2002 and 2008 [5–8]) had shownthat antenatal MgSO4 administered for fetal neuropro-tection (NP) at < 32–34 weeks reduces the likelihood ofCP (relative risk (RR) 0.68 [0.54, 0.87]) [9–11]. However,controversies remained about this therapy, includingconcerns about potential effects of MgSO4 on fetal heartrate [12] and increased neonatal resuscitation [13], a lackof understanding of the neuroprotective mechanism ofaction [14] and inadequate studies describing long-termadverse paediatric outcomes other than CP.RationaleAs antenatal corticosteroids prior to preterm delivery werenot routinely administered in North America until 22 yearsafter their benefit had been established, we anticipated thatimplementation of MgSO4 for fetal NP into clinical prac-tice would require a knowledge translation (KT) interven-tion. A previous study on existing knowledge resourcesabout MgSO4 for fetal NP in Canada found that despiteconvincing evidence of effectiveness, use of MgSO4 for fetalNP was near non-existent (1.5%) between 2010 and 2011,and there was no such use of MgSO4 before 2010 [15]. Still,knowledge gaps and lack of guidelines remained importantbarriers to use [16, 17], with the potential to cause maternalside effects an additional anticipated barrier as it has beenfor implementation of MgSO4 for eclampsia preventionand treatment [18]. As maternity care hospitals vary widelyin terms of practices and beliefs, MAG-CP (MAGnesiumsulphate for fetal neuroprotection to prevent CerebralPalsy) was created to facilitate uptake of use of MgSO4 forfetal NP in the setting of imminent birth at < 32 weeks.KT strategiesWe chose a multifaceted implementation strategy thatwas informed by the concepts of Roger’s Innovation-Diffusion theory [19], the most influential theory inknowledge utilisation [20]. This theory considers thecomplexity of the innovation or clinical practice, charac-teristics of adopters, communication channels, time con-siderations for adoption and uptake and organisationalcharacteristics of the social system [19]. The process ofbehaviour change at the individual level includes know-ledge of the innovation or clinical practice, persuasion foruptake, an individual decision for uptake and use at whichpoint the innovation is either accepted or rejected, imple-mentation of the innovation or clinical practice and con-firmation of the decision for uptake [19]. We specificallyincluded e-learning platforms and site outreach activitiesshown to support active (rather than passive) learning (es-pecially when those activities are used in conjunction withother interventions [21, 22]) and audit and feedback thathave been effective in improving practice [23–25].ObjectivesOur primary aim was to describe our multifaceted imple-mentation strategy and assess its effectiveness in increas-ing ‘optimal’ use of MgSO4 (i.e. MgSO4 administration towomen delivering at under 32 weeks as indicated and nouse when not indicated) to 80% of eligible women over4 years (2011–2015), the standard benchmark for a grade1A recommendation [26]. Our secondary objective was toreport any maternal or fetal adverse effects of our healthintervention given the importance of such effects in theimplementation process. We describe our KT strategy andtargeted sites, outcome measurement and data analysisusing data from the Canadian Perinatal Network (CPN)and Canadian Neonatal Network (CNN).MethodsKT strategy (2011–2015) and targeted centresWe undertook an interrupted time-series study designusing segmented regression analysis to evaluate the effect-iveness of a selected bundle of KT strategies to optimise useof MgSO4 for fetal NP. We have employed the use of theStaRI (Standards for Reporting Implementation Studies) asour reporting standard [27].The strategy consisted of four parts: (1) initiating andleading a Society of Obstetricians and Gynaecologists ofCanada (SOGC) clinical practice guideline on the topicthat was published in May 2011 [26] and then from 2011to 2015 (as previously detailed [16]), (2) an e-learningmodule; (3) a ‘Barriers and Facilitators Survey’ and (4) anaudit and feedback cycle, including site visits, monitoringDe Silva et al. Implementation Science  (2018) 13:8 Page 2 of 16and other interactive activities between the central MAG-CP team and individual sites. We have previously pub-lished a qualitative analysis of our strategies [16] (Fig. 1).The first two components of the KT strategy were avail-able to all practitioners in Canada. The SOGC guidelineswere open-access and free to anyone who received theSociety’s journal or had internet access. The e-learningmodule was created by the central MAG-CP team, pub-lished by AdvancingIn®, an online platform providing elec-tronic continuing medical education, which was free to allCanadian health care professionals who were SOGCmembers. The module was developed by the seniorauthors of the guideline on behalf of the SOGC, who sep-arately contracted AdvancingIn® for their platform. Thismodule included pre-test questions, a concise summary ofthe evidence and the 2011 SOGC guideline, a summary ofcontroversies and uncertainties, case analyses, practicetools, post-test questions, and a discussion forum regardingMgSO4 for NP. The module was incentivised by providingcontinuing medical education (CME) credits to eligiblehealth care professionals and a certificate of completion.The third and fourth components of the KT were deliv-ered to practitioners at participating study sites in theCanadian Perinatal Network (CPN) that agreed to partici-pate in the KT activities and collect relevant outcome data(see below). These centres were tertiary perinatal centresthat were likely to see women eligible for the intervention.The ‘Barriers and Facilitators (B&F) survey’, informed bythe Theoretical Domains Framework [28], was distributedat each MAG-CP study participating site, by each site’slocal team to be completed by at least five obstetriciansand five nurses (to explore local barriers to and facilitatorsof MAG-CP implementation). The surveys were anon-ymised and consisted of mixed free text and tick choices,collected locally, and sent to the central team for compil-ation and interpretation to provide feedback of results toeach site for their review. This approach was chosen todetermine organisation readiness and address challengesas well as identify knowledge gaps and tailor interventions.Further details and a copy of the survey have been pub-lished in our detailed qualitative analysis [16].The audit and feedback cycle to address local barriers con-sisted of visits to study centres that were organised by thecentral team. They presented didactic grand rounds andfacilitated small site-specific interactive group discussion.Other activities for feedback and exploration of barriersincluded a monthly newsletter, monthly teleconferences,supportive emails and one-on-one support for questionsand advice and provision of KT tools (presentation mate-rials, information sheets for staff and women and remindersfor women who were being expectantly managed in hospitaland at risk of preterm birth at < 32 weeks [29]). As part ofSOGC Guidelines*Provide feedbackǂCentral MAG-CP Team and siteMonitor use of MgSO4Monitor clinical outcomesFor MAG-CP centrese-learning module* Barriers & Facilitators surveyFig. 1 Schematic of the MAG-CP knowledge translation audit cycle. *All members of the Society of Obstetricians and Gynaecologists of Canada(SOGC) were sent the guideline and link to the e-learning module, both of which were open-access to anyone else who was made aware ofthem. †Central MAG-CP team interactions with each site included site visits to study centres where members of the central team presented didactic grandrounds and facilitated small site-specific interactive group discussion; a monthly newsletter; monthly teleconferences; supportive emails and one-on-onesupport for questions and advice; provision of KT tools (such as pre-printed physician orders, presentation materials, information sheets for staff andwomen; and reminders for women who were being expectantly managed in hospital and at risk of preterm birth at < 32 weeks [www.cpn-rpc.org]).ǂFeedback included semi-annual site-specific reports on MgSO4 for fetal NP use that compared each site with activity overall (while maintaining theanonymity of other sites), thus creating an audit cycle to inform and fuel ongoing KTDe Silva et al. Implementation Science  (2018) 13:8 Page 3 of 16the feedback, the central MAG-CP team prepared 6-monthsite-specific reports on MgSO4 for fetal NP use that com-pared each site with activity overall (while maintaining theanonymity of other sites), thus creating the audit cycle toinform and fuel ongoing KT.Interventions by the central MAG-CP team (such as e-learning module, site visits, completeness of B&F surveysand number of teleconferences participated) were directlymeasured as a form of fidelity of the KT strategies, whilesites were asked to record other local KT strategies (thatreflected participant responsiveness and other potentialmoderators of the intervention-adherence relationship[30]) using a web-based form (such as use of decision-aidtools, reminders, presentations or ‘teaching moments’).This was summarised as an ‘engagement’ measure of siteparticipation in the intervention (Additional file 1: TableS7). ‘Highly engaged’ sites had values that were above themedian or mean for each activity and overall; this assess-ment was conducted independently by each member ofthe MAG-CP working group (DAD, LAM and ARS), eachof whom was masked to site identity. Discrepancies wereresolved by consensus.Evaluation of KT strategies via the CPNEnrollment into the CPNTo evaluate the effectiveness of the KT strategy, wetracked MgSO4 use and outcomes using data from theCanadian Perinatal Network (CPN) [31]. In brief, theCPN, another CIHR-funded project, collected demo-graphic, management and outcome information (August2005 and September 2015, inclusive) on women admit-ted at 22 weeks (and 0 days) to 28 weeks (and 6 days)with threatened very preterm birth to participating ter-tiary perinatal centres and who were followed to deliv-ery. These women were admitted with one or more ofspontaneous preterm labour with contractions, pretermpre-labour rupture of membranes (PPROM), short cer-vix without contractions, prolapsing membranes, gesta-tional hypertension, intrauterine fetal growth restriction(IUGR) or antepartum haemorrhage (see Additional file 1:Table S2 for definitions). The project was approved cen-trally by the Research Ethics Board at the University ofBritish Columbia (H05-70359 and H11-02214) andlocally at each participating centre. As this was approvedas a quality improvement project with no patient con-tact, patient consent was not deemed necessary.Sample size calculationOver the 4 years of the KT strategy, we anticipated recruit-ment of 3752 mothers based on previous CPN enrollmentof women at < 29 weeks (from the CPN inception in2005). We estimated that we would have > 95% power(two-sided alpha of 0.05, < 5% baseline use of MgSO4 forfetal NP) for each of two scenarios: (i) ‘planned’ rates ofMgSO4 use for fetal NP of 20, 40, 60 and 80% by the endof years 1–4, respectively, and (ii) ‘pessimistic’ rates of 20,30, 40 and 50% by the end of years 1–4, respectively, basedon a prior survey with the centres. The power calculationswere made without adjustment for random effects (i.e.clustering), because the calculations for these adjustmentsalso require specification of the distribution of MgSO4use across hospitals in the 4-year study period, and thesewere not known.For adverse maternal outcomes, we estimated at least80% power to detect potential increases in serious maternaladverse effects reported in RCTs: hypotension (RR 1.51[1.09, 2.09] from baseline of 6.5%), infusion stopped due toadverse effects (RR 2.81 [2.01, 3.93] from 2.6%), respiratorydepression (RR 1.31 [0.83, 2.07] from 1.9%) and pulmonaryoedema (RR 2.79 [0.74, 10.47] from 0.3%) [26].No increase or decrease in stillbirth or neonatal deathwas anticipated, but we were powered to detect only sub-stantial increases in these outcomes (i.e. an increase of27–28% in total paediatric mortality under the plannedand ‘pessimistic’ pre-specified rates of MgSO4 use).Sampling of eligible casesPregnancies were tracked in participating CPN sites in‘pre-’ and ‘post-’ MAG-CP eras (dates inclusive): (i) ‘pre’-MAG-CP (‘controls’, August 2005 to May 2011) and (ii)‘post’, termed MAG-CP (‘cases’, June 2011 to September2015). The pre-MAG-CP era represented the period be-fore the KT intervention, from (i) the beginning of CPNdata collection through the publication of the last pri-mary trial of MgSO4 for fetal NP (2005–2008) [5–8], (ii)the publication of three independent systematic reviewsof those primary trials (2009) [9–11] and (iii) the periodthereafter until publication of the Canadian SOGC Clin-ical Practice Guidelines on MgSO4 for fetal NP [26] inMay 2011 (Jan 2010–May 2011). Data were includedfrom all CPN sites. The MAG-CP era ran from June2011 until Sept 2015 (sub-divided into nine 6-monthtime periods) and included data from CPN sites thatchose to participate in the MAG-CP study.CPN-eligible pregnancies were ≥ 24 weeks (and 0 days)at CPN enrollment, presented with imminent pretermbirth (i.e. likely within 24 h) at < 32 weeks (and 0 days)(using criteria consistent with the relevant primary trialsand as summarised in Canadian guidance [26]) and werefollowed to delivery. CPN pregnancies were excluded ifwomen had received MgSO4 for an indication otherthan fetal NP (e.g. eclampsia prophylaxis or treatment).Data collection and analysis of outcomeMaternal information collected from CPN included mater-nal characteristics, obstetric history, details of hospital ad-mission, maternal and fetal surveillance, labour and delivery,maternal outcomes, other maternal interventions (includingDe Silva et al. Implementation Science  (2018) 13:8 Page 4 of 16MgSO4 administration and indication), stillbirth and neo-natal outcome (other than resuscitation which was not avail-able in the CPN). As the CPN database was revised forMAG-CP to include detailed information on MgSO4administration (including dose, duration and side effects),this information was only available for pregnancies duringthe MAG-CP era (2011–2015).The primary outcome was the rate of optimal MgSO4use (i.e. administration when and only when needed), overtime. ‘Underuse’ was defined as failure to administerMgSO4 for fetal NP when indicated (i.e. for birth that oc-curred within 24 h of admission to hospital at < 32 weeks(and 0 days)) and suboptimal use as administration ofMgSO4 for fetal NP when not indicated, either when birthdid not occur within 24 h at < 32 weeks (and 0 days) orbirth occurred at ≥ 32 weeks (and 0 days). Secondary out-comes monitored included adverse maternal and neonataleffects. Maternal adverse effects included hypotension (i.e.diastolic blood pressure fall of > 15 mmHg), the need tostop MgSO4 because of side effects (‘stopped infusion’),respiratory depression (i.e. < 12 breaths/min) and pulmon-ary oedema (as per the clinician’s assessment).Descriptive statistics were used to summarise maternalcharacteristics, details of admission and outcomes, withchi-square or Mann-Whitney U test used where appropri-ate. As the primary outcome was the monthly rate of opti-mal MgSO4 use over time, we performed an interruptedtime-series analysis, a powerful quasi-experimental studydesign, to evaluate the effect of the KT intervention inMAG-CP compared with pre-MAG-CP eras and distin-guish it from any observed effects in the absence of inter-vention [32, 33]; segmented generalised estimatingequations (GEE) logistic regression was used to accountfor centre variability. Also, MgSO4 use for fetal NP wascompared between ‘highly engaged’ and ‘less engaged’sites. To correct for pregnancies that may have been pre-cipitous in nature, we adjusted for any administration ofantenatal corticosteroids, reasoning that there would beenough time to administer MgSO4 for fetal NP if ante-natal corticosteroids were administered. Sensitivity ana-lyses were conducted using data only from centres thatparticipated in both the pre- and MAG-CP eras to assessusage rates of MgSO4. All statistical analyses were per-formed using R statistical software [34]. A p value < 0.05was considered statistically significant.Canadian Neonatal Network (CNN) data collection andanalysisTo further explore trends in MgSO4 use for fetal NP inCanada and associated neonatal resuscitation rates, weobtained data from the CNN that collects information onbabies admitted to neonatal intensive care units (NICUs)in 31 participating NICUs in Canada. We included babiesborn at 24 weeks (and 0 days) to 31 weeks (and 6 days)and collected information about use of MgSO4 for fetalNP as well as pregnancy characteristics. We examined theproportion of babies who received MgSO4 for fetal NPfrom January 2011 when ‘fetal NP’ was first listed as anindication for MgSO4 in the CNN database; as such, datawere available for a portion of the pre-MAG-CP (January–May 2011, inclusive) and during MAG-CP (i.e. June 2011–September 2015, inclusive). Also, we examined rates ofintensive neonatal resuscitation, defined by the CNN as theneed for either (i) chest compressions or intubation andventilation or (ii) epinephrine administration in the deliveryroom [35]. GEE logistic regression was used to examinewhether MgSO4 use for fetal NP changed over time.Among babies exposed to MgSO4 for fetal NP (comparedwith those who did not receive MgSO4 or received it foran indication other than fetal NP), logistic regression wasused to calculate the odds ratio (OR) for ‘intensive’neonatal resuscitation. GEE was used to adjust for import-ant covariates (i.e. multiple gestation, gender, gestationalage at delivery, birth weight < 10th centile, outborn status,mode of delivery and antenatal corticosteroid use), andbabies with congenital anomalies were excluded, as inprior CNN analyses [35]. A p value < 0.05 was consideredstatistically significant.Role of the funding sourceThe funder of the study had no role in study design, datacollection, data analysis, data interpretation or writing ofthe report. The corresponding author as well as DAD andARS had full access to all the data in the study and had finalresponsibility for the decision to submit for publication.ResultsParticipating centresEighteen of Canada’s 23 tertiary perinatal centres contrib-uted data to either the pre-MAG-CP or MAG-CP eras;nine centres contributed outcome data continuously fromAugust 2005–September 2015, seven contributed onlypre-MAG-CP and two contributed only during MAG-CP.Thus, eleven centres received the KT strategy from 2011to 2015. Participating centres were from all geographic re-gions of Canada (i.e. 6 Western centres, 8 Ontario/Quebeccentres and 4 Atlantic centres), with annual delivery vol-umes ranging from < 2000 to ≥ 5000. For details of siteparticipation, see Additional file 1: Table S3.Implementation fidelity was variable between sites, asreflected in our measure of ‘engagement’ with KT. Ten of11 (90.9%) centres completed the e-learning module, andall centres completed the B&F surveys with a median of15 respondents, as well as receiving audit and feedback.Eight centres received a site visit, covering 10/11 MAG-CP sites (Additional file 1: Table S7). One site visit couldnot be arranged at a mutually convenient time. CentresDe Silva et al. Implementation Science  (2018) 13:8 Page 5 of 16did not appear to be exposed or respond to the KT strat-egies equally.Outcomes—Canadian Perinatal Network dataThere were 5683 women enrolled at 18 CPN sites during thepre-MAG-CP era, of whom 4745 (83.5%) were included foranalyses. Similarly, there were 3784 women enrolled at 11CPN sites (participating in MAG-CP) during the MAG-CPera, of whom 3143 (83.1%) were included. The proportion ofeligible patients did not differ between the pre-MAG-CP andMAG-CP eras (p= 0.60).Characteristics of the sampleThere were differences in the characteristics of womenenrolled in pre-MAG-CP and MAG-CP eras (Table 1,which also presents maternal and perinatal outcomes forcompleteness). Many differences were small in magnitude(e.g. history of venous thromboembolism) and/or of ques-tionable clinical significance (e.g. maternal age and gesta-tional age at enrollment in CPN and at delivery). Overall,women were just over 30 years of age. Few women (< 5%)had pre-existing medical conditions. Among parouswomen, about one-third had experienced prior pretermdelivery. Approximately half of women were nulliparousand almost 20% had multiple pregnancies. Most womenwere non-smokers in the current pregnancy, particularlyduring the MAG-CP era. Women were enrolled in CPNat about 26 weeks, usually for preterm labour and/orPPROM, and they delivered at about 30 weeks’ gestation.In terms of other maternal and perinatal outcomes, preg-nancies in the MAG-CP era were more often complicatedby abruption and serious maternal complications,although stillbirth and neonatal death were less frequent.Analysis of MgSO4 usageMgSO4 for fetal NP was administered (either ‘optimally’to those who needed it or ‘suboptimally’ to those whodid not need it, as previously defined, see the ‘Methods’section) to 94 (2.0%) of women in the pre-MAG-CP eraand 1454 (46.3%) during MAG-CP. Details of MgSO4administration were collected only during MAG-CP.During the MAG-CP era, women received MgSO4 forfetal NP at about 27 weeks, approximately 1 week afteradmission with threatened very preterm birth (Table 2).Almost 40% of women were in active labour with ≥ 4 cmof cervical dilatation at the time of receiving MgSO4.More than 90% of women received one course ofMgSO4 for fetal NP, with 82.4% of women receivingboth loading dose and maintenance doses. The usualloading dose was 4 g iv over a median of 25 min. Themedian maintenance dose was 1 g/h iv for 7.4 h. MgSO4was associated with few adverse effects. There were eightepisodes of hypotension among seven women (0.5%) fol-lowing loading (N = 5 episodes) or maintenance dosing(N = 3); none of the seven women had their infusionsdecreased or stopped or calcium gluconate administered.No woman experienced respiratory depression. Elevenwomen (0.7%) experienced pulmonary oedema. Otherside effects (unspecified) occurred in one additionalwoman (0.1%) who was given calcium gluconate. Mostwomen had their MgSO4 infusions stopped because theyeither delivered or were no longer considered at risk ofimminent preterm birth.Segmented regression analysisTable 3 presents the odds ratios of MgSO4 for fetal NP inthe pre-MAG-CP and MAG-CP eras, according to opti-mal, under- and suboptimal use among eligible women.The absolute rates are presented in Additional file 1: TableS4, including the number of women in the optimal usecategory who needed MgSO4 for fetal NP and got it. Inthe pre-MAG-CP era, the odds of optimal use wereincreasing by 0.4% (OR 1.004 [0.997–1.01]) per month.Upon the start of the KT intervention, there was an im-mediate 84% increase in the odds of optimal use (OR 1.84[1.51–2.24]), after which, there was a significant continu-ous increase of 2% (OR 1.02 [1.00, 1.04]) per month,compared to the pre-MAG-CP era (p < 0.001) (Fig. 2a).Thus, there was a 220% increase in the odds of optimaluse during the MAG-CP era (OR 3.20); by comparison,the anticipated increase in optimal use, assuming that theKT intervention had not occurred, would be only 23%(OR 1.23). This increase in optimal use is mirrored by afall in underuse.The initial optimal use rate (36.0%) was related tonon-administration to women for whom MgSO4 forfetal NP was not indicated (i.e. 751/2088, 36.0%), ratherthan administration to women for whom MgSO4 forfetal NP was indicated (i.e. 0/2088, 0%) (Fig. 3a, b). Theodds of administration of MgSO4 for fetal NP to eligiblewomen, termed ‘appropriate’ use, significantly increasedupon the start of the KT intervention (p < 0.001) (Fig. 2b)and within the MAG-CP era specifically (p < 0.001).Suboptimal use of MgSO4 for fetal NP also increasedimmediately after the KT intervention (OR 2.18, p = 0.038;Table 3) and continued to increase per month in theMAG-CP era but at a slower rate than in the pre-MAG-CPera. However, absolute rates remained quite low (< 13%)throughout the study (Additional file 1: Table S4). Althoughthe SOGC guideline recommended only in the text (ratherthan in the recommendations) to consider MgSO4 for fetalNP for women between 32+0 and 33+6 weeks’ gestation withimminent preterm birth, few such women (79/3143, 2.5%)received such treatment during the MAG-CP era.Sensitivity analyses based on the nine centres (7066women) that contributed data to both pre-MAG-CPand MAG-CP eras were similar to the overall results(Additional file 1: Table S5 and S6).De Silva et al. Implementation Science  (2018) 13:8 Page 6 of 16Table 1 Baseline characteristics and pregnancy outcomes of women with imminent preterm birth at participating Canadian PerinatalNetwork sites (2005–15) (N (%) women or median [IQR], unless otherwise stated)TotalN = 7888Pre-MAG-CP (2005–2011)N = 4745MAG-CP (2011–2015)N = 3143p valuesMaternal demographics and past historyMaternal age at EDD (year) 31 [27, 35] 31 [27,35] 31 [27,35] 0.034Pre-existing medical conditionsPre-existing hypertension 305 (3.9%) 178 (3.8%) 127 (4.0%) 0.600Diabetes mellitus 149 (1.9%) 79 (1.7%) 70 (2.2%) 0.090Venous thromboembolism 28 (0.4%) 25 (0.5%) 3 (0.1%) 0.001Prior obstetric historyPrevious preterm birth 1383 (17.5%) 845 (17.8%) 538 (17.1%) 0.400Previous caesarean 930 (11.8%) 666 (14.0%) 264 (8.4%) < 0.001Current pregnancyNulliparity 3909 (49.6%) 2303 (48.5%) 1606 (51.1%) 0.030Multiple gestation 1507 (19.1%) 911 (19.2%) 596 (19.0%) 0.800Smoking during pregnancy 1260 (16.0%) 833 (17.6%) 427 (13.6%) < 0.001Missing 52 33 19Gestational age at enrollment (week) 26.0 [24.4, 27.4] 26.1 [24.6, 27.6] 25.9 [24.4, 27.4] < 0.001Indication for threatened preterm birthPreterm labour only 2324 (29.5%) 1375 (29.0%) 949 (30.2%) < 0.001PPROM only 1567 (19.9%) 960 (20.2%) 607 (19.9%)PTL and PPROM 1106 (14.0%) 591 (12.5%) 515 (16.4%)Antepartum haemorrhage only 1195 (15.1%) 764 (16.1%) 431 (13.7%)Other† 1696 (21.5%) 1055 (22.2%) 641 (20.4%)Gestational age at delivery (week) 28.0 [26.0, 35.0] 28.0 [26.0, 35.0] 28.0 [26.0, 34.0] 0.036≥ 37 weeks (and 0 days) 1404 (17.8%) 889 (18.7%) 515 (16.4%) 0.00834 weeks (and 0 days)–36 weeks (and 6 days) 821 (10.4%) 507 (10.7%) 314 (10.0%)29 weeks (and 0 days)–33 weeks (and 6 days) 1522 (19.3%) 926 (19.5%) 596 (19.0%)< 29 weeks (and 0 days) 4141 (52.5%) 2423 (51.1%) 1718 (54.7%)Maternal outcomesPlacental abruption after enrollment 635 (8.1%) 251 (5.3%) 384 (12.2%) < 0.001One/more serious maternal complications 2479 (31.4%) 1304 (27.5%) 1175 (37.4%) < 0.001Death 2 (0.03%) 2 (0.04%) 0 0.500Admission to ICU or HDU 98 (1.2%) 19 (0.4%) 79 (2.5%) < 0.001Chorioamnionitis 1789 (22.7%) 999 (21.1%) 790 (25.1%) < 0.001Cardiovascular 2 (0.03%) 1 (0.02%) 1 (0.03%) 0.999Respiratory 64 (0.8%) 38 (0.8%) 26 (0.8%) 0.999CNS 7 (0.09%) 5 (0.1%) 2 (0.06%) 0.700Renal 6 (0.08%) 4 (0.08%) 2 (0.06%) 0.999Hematological 72 (0.9%) 37 (0.8%) 17 (0.5%) 0.300Hepatic 11 (0.1%) 11 (0.2%) 0 0.004Infection 74 (0.9%) 50 (1.1%) 24 (0.8%) 0.200Perinatal outcomes N = 9541 N = 5751 N = 3790Stillbirth 291 (3.0%) 211 (3.7%) 80 (2.1%) < 0.001Neonatal death in the delivery room 164 (1.7%) 119 (2.1%) 45 (1.2%) 0.002Liveborn and admitted to NICU 7638 (80.1%) 4714 (82.0%) 2924 (77.2%) <0.001PPROM preterm premature rupture of membranes, CNS central nervous system, HDU high-dependency unit, ICU intensive care unit, IQR interquartile range, NICU neonatalintensive care unit†Other indications for threatened preterm birth in the absence of preterm labour, PPROM or antepartum haemorrhage included (not mutually exclusive) gestationalhypertension (N = 227), intrauterine growth restriction (N = 238), short cervix (N = 476), prolapsed membranes (N = 249) or other non-CPN condition within the Maternal-InfantCare Network (N = 48) in the pre-MAG-CP era and gestational hypertension (N = 137), intrauterine growth restriction (N = 151), short cervix (N = 325) or prolapsed membranes(N = 157) in the MAG-CP eraDe Silva et al. Implementation Science  (2018) 13:8 Page 7 of 16Table 2 Details of 1512 women who received MgSO4 for fetal NP administration and adverse effects during MAG-CP (N (%) womenunless otherwise specified)N (%) or median [IQR]Gestational age at time of MgSO4 therapy (week) 27.1 [25.6, 28.4]Cervical dilatation at time of therapy (cm) 3 [1.5,4.5]Cervical dilatation ≥ 4 cm 570 (37.7%)Cervical dilatation ≥ 4 cm among women with PTL 384/728 (52.7%)Missing/unknown 227 (15.0%)N treatment courses/woman 1 [1,1]Received more than one course 110 (7.3%)Received only loading dose 222 (14.7%)Received only maintenance dose 44 (2.9%)Received both loading and maintenance doses 1246 (82.4%)Loading dose detailsRoute of administrationIV only 1464 (96.8%)IM only 2 (0.1%)Initial dose (g) 4 [4,4]Duration of therapy (min) 25 [20, 30]Missing 95 (6.5%)Adverse maternal effects (one/more) 6 (0.4%)Maternal hypotension 5 (0.3%)Respiratory depression 0Pulmonary oedema 1 (0.1%)Loading dose stopped early 45 (3.1%)Stopped because woman delivered 38 (2.6%)Stopped because patient refused treatment or further treatment 1 (0.1%)Stopped because of maternal side effects 0Stopped because woman was no longer in imminent preterm birth 1 (0.1%)Other* 4 (0.3%)Calcium gluconate administered 0Maintenance dose detailsRoute of administrationIV only 1290 (85.3%)IM only 0Initial dose (g/h) 1 [1,1]Duration of therapy (h) 7.4 [3.1, 17.5]Adverse maternal effects (one/more) 13 (0.9%)Maternal hypotension 3 (0.2%)Respiratory depression 0Pulmonary oedema 10 (0.7%)Reasons for stopping maintenance doseStopped because woman delivered 738 (57.2%)Stopped because 24 h of therapy had been administered 81 (6.3%)Stopped because woman was no longer in imminent preterm birth 163 (12.6%)Stopped because of maternal side effects 0De Silva et al. Implementation Science  (2018) 13:8 Page 8 of 16Despite the strategies, there was substantial between-centre variability in optimal use and underuse rates ofMgSO4 for fetal NP (Fig. 4a, b); two sites had optimaluse rates ≥ 95th centile and four sites rates ≤ 5th centile.In general, sites with high optimal rates had lowerabsolute underuse rates and vice versa. There was farless variability seen in suboptimal use (Fig 4c); threeof 11 sites had suboptimal use rates that were ≥ 95thcentile, most often related to use among women whodid not deliver imminently (i.e. within 24 h, 213/3143[6.8%]) (Additional file 1: Table S4).Site engagementSeven of 11 MAG-CP sites were ranked as ‘highlyengaged’ and four as ‘less engaged’, based on criterialisted in Additional file 1: Table S7. However, neitherthis overall measure of engagement, nor the individ-ual components of KT on which the assessment wasbased, including those directly measurable by the cen-tral team and reflecting implementation fidelity, wereassociated with optimal use ≥ 95th centile, althoughthere appeared to be a strong trend towards suchhigh optimal use and early ‘buy-in’ by participation inMAG-CP data collection (Additional file 1: Table S8).Canadian Neonatal Network (CNN) dataAmong 14,108 infants born at 24–31 completedweeks and admitted to NICU in 31 CNN sites (Janu-ary 2011–September 2015), there was a significant in-crease in use of MgSO4 for fetal NP over time, from19.7% pre-MAG-CP (Jan 01, 2011 to May 31, 2011)to 62.4% in the last MAG-CP time period (Apr 1,2015 to Sept 30, 2015) (p < 0.001; Additional file 1:Table S9). There was, however, substantial between-centre variability in the use of MgSO4 for fetal NP,and optimal use ≥ 99th percentile was more frequentat MAG-CP sites (8/11) than at other CNN sites (4/19, p = 0.015; Additional file 1: Figure S3). No sitesdemonstrated optimal use that was between the 5thand 95th percentiles.Antenatal and birth characteristics of infants dif-fered between infants who were either unexposed,exposed for fetal NP or exposed for other indications(Table 4). Following adjustment for multiple gestation,gender, gestational age at delivery, birth weight < 10thcentile, outborn status, mode of delivery and ante-natal corticosteroid use, MgSO4 for fetal NP wasassociated with a lower risk of intensive neonatalresuscitation compared with either (i) non-receipt ofMgSO4 (adjusted OR 0.63 [0.54, 0.73], p < 0.001) or(ii) receipt of MgSO4 for an indication other thanfetal NP (adjusted OR 0.81 [0.66, 0.99], p = 0.04).DiscussionSummary of resultsIn an interrupted time-series analysis of a large cohortof women who were enrolled in the CPN followingadmission with threatened preterm birth and who wereeligible for MgSO4 for fetal NP, we found that ourmultifaceted KT strategy resulted in a significantincrease in optimal use of MgSO4 for fetal NP. This wasTable 2 Details of 1512 women who received MgSO4 for fetal NP administration and adverse effects during MAG-CP (N (%) womenunless otherwise specified) (Continued)N (%) or median [IQR]Other† 29 (2.2%)No reason indicated or missing 279 (21.6%)Calcium gluconate administered 1 (0.1%)CPN Canadian Perinatal Network*Other reasons for stopping the loading dose of MgSO4 for fetal NP early were emergency caesarean (N = 1), patient in extreme pain from IV (N = 1), patient feltburning/flushing (N = 1) and unknown (N = 1)†Other reasons for stopping the maintenance dose of MgSO4 for fetal NP were as per protocol or other orders (e.g. 12 h of therapy administered) (N = 9), dosagechange (N = 7), patient transferred (N = 4), MgSO4 continued postpartum for pre-eclampsia prevention (N = 3), fetal demise (N = 2), emergency caesarean (N= 2) orpatient experienced side effects (N = 2)Table 3 Overall odds ratios for use of MgSO4 for fetal NP as derived from segmented regression analysisOptimal use* p value Underuse* p value Suboptimal use* p valueOdds ratio for use in pre-MAG-CP, per month† 1.004 [0.997, 1.01] 0.226 0.995 [0.99, 1.00] 0.104 1.18 [1.08, 1.28] < 0.001Immediate change in odds just after intervention† 1.84 [1.51, 2.24] < 0.001 0.47 [0.34, 0.65] < 0.001 2.18 [1.04, 4.58] 0.038Change in odds ratio after intervention compared to pre-MAG-CP,per month†1.02 [1.00, 1.04] 0.044 0.97 [0.95, 0.99] 0.002 0.86 [0.79, 0.94] < 0.001Odds ratio for use in MAG-CP era, per month† 1.02 [1.01, 1.03] < 0.001 0.97 [0.95, 0.98] < 0.001 1.01 [1.001, 1.02] 0.027*Optimal use refers to both women who received MgSO4 for fetal NP when indicated, as well as women who did not receive MgSO4 for fetal NP when it was notindicated. Underuse refers to eligible women who should have received MgSO4 for fetal NP but did not. Suboptimal use refers to women who received MgSO4too early (not within 24 h before birth) or at ≥ 32 weeks. †Segmented regression analysis was adjusted for antenatal administration of corticosteroidsDe Silva et al. Implementation Science  (2018) 13:8 Page 9 of 16primarily due to an increase in administration towomen who were eligible to receive it, with a min-imal increase in administration to women who wenton to deliver more than 24 h later or at ≥ 32 weeks(and 0 days). Although women in the pre-MAG-CPand MAG-CP eras differed according to a number ofmaternal and pregnancy characteristics, none wouldinfluence administration of MgSO4 for fetal NP andcould be expected to account for the trends observedin MgSO4 use, particularly as these differences overtime were observed only during the MAG-CP era.In addition, using data from the CNN, tertiary perinatalcentres that participated in MAG-CP (compared with thosethat did not) had higher optimal use rates of MgSO4 forabFig. 2 Segmented regression analysis of pre-MAG-CP (2005–11) and MAG-CP (2011–2015) eras. a Optimal use. b Appropriate use. The dashed lineindicates implementation of the KT interventionDe Silva et al. Implementation Science  (2018) 13:8 Page 10 of 16fetal NP. Finally, MgSO4 for fetal NP was safe, with few ma-ternal side effects (documented in MAG-CP centre) despitewomen having pregnancies more frequently complicated byplacental abruption and other maternal complications dur-ing the MAG-CP era and no increase in neonatal resuscita-tion (demonstrated in CNN sites).abFig. 3 Absolute utilisation rates over pre-MAG-CP (2005–11) and MAG-CP (2011–2015) eras. a Optimal use. b Appropriate use. The solid black lineindicates the median overall rate and each coloured line represents one of the 11 participating centresDe Silva et al. Implementation Science  (2018) 13:8 Page 11 of 16How the findings fit with the published literatureKT approachMultifaceted KT approaches can be more effective thandissemination alone in encouraging the adoption andimplementation of new research results [36], changingclinical outcomes [37–40] and achieving improvementsin policy and practice [41]. Specifically, audit and feed-back are effective enablers of evidence-based guidelineimplementation [42]. While older reviews foundmultiple KT interventions to be more effective thansingle-strategy approaches [43–45], this is not necessar-ily the case in more recent literature in which singleinterventions can have an impact similar to multifacetedapproaches [46]. However, this may vary according tothe circumstances, such as the complexity of the healthintervention or the organisational culture in which theintervention is implemented [19, 36]. This was what wefound in a published analysis of the relative merits ofour online e-learning module, interactive site visits (witheducational rounds and focus group discussions) andcirculation of an anonymous Barriers and Facilitatorssurvey to systematically identify barriers to and facilita-tors of practice change [16]. In brief, no individual KTmethod was superior to the others with regards to (i)Fig. 4 Between-centre variability of MgSO4 usage among centres. a Variability in optimal use. b Variability in underuse. c Variability in suboptimaluse. The solid line indicates 95% confidence interval while the dotted line indicates the 99% confidence intervalDe Silva et al. Implementation Science  (2018) 13:8 Page 12 of 16breadth of respondents reached; (ii) rates and richness ofidentified barriers, facilitators, and knowledge needed;and (iii) cost, in combination. The e-learning modulereached the most diverse audience of health care pro-viders, the site visits provided opportunity for iterativedialogue and the survey was the least expensive.Although the site visits provided the most detailed infor-mation around individual and organisational barriers,the ‘Barriers and Facilitators’ survey provided more detailabout social-level barriers.We recognise that costs and resources have implica-tions. The bulk of our resources were used in the collec-tion of outcome data. If health care systems routinelycollected information related to monitoring the out-come, the cost of the project would consist of creationof resources and support of the local KT teams to moveit forward.Health interventionMany international societies and bodies have now issued clin-ical practice guidelines that recommend MgSO4 for fetal NPin the setting of imminent preterm birth at < 32–34 weeks[26, 47–49]. The implementation of MgSO4 for fetal NP hasbeen evaluated in single centres in the USA (that led theBEAM trial) [50], France [51], New Zealand [52] andAustralia, where a similar multicentre KT implementationproject is ongoing [53]. Although final analyses and follow-upresults of the latter study are still pending, preliminary resultsin one institution (where the Australian primary trial wasundertaken) showed a fall in underuse rates from 69.7 to26.9% over 2 years [54]. Although we are not aware of othersuch KT initiatives in relation to MgSO4, the need for themhas been recognised [55]. Studies of international practiceconfirm practice heterogeneity [17, 56, 57]).Historically, MgSO4 has been regarded as increasingthe risk of neonatal respiratory depression, hypotoniaand the need for resuscitation [58]. However, our findingthat use of MgSO4 for fetal NP does not increase (butrather is associated with a decrease in) the need forintensive neonatal resuscitation at delivery is consistentwith more recent literature that has demonstrated noincrease in resuscitation [13, 59–61] and, in some cases,a decreased need [35]. To date, no adverse effects ofMgSO4 for fetal NP have been demonstrated on fetalTable 4 Selected infant characteristics and outcomes according to exposure to MgSO4 for fetal NP and its indication (N (%) ormedian [IQR], where appropriate, unless otherwise indicated)MgSO4 for neuroprotectionN = 5314No MgSO4N = 7238MgSO4 for another indicationN = 1556p valueAntenatal and birth characteristicsGA, weeks and days 29 [26, 30] 29 [27, 30] 29 [27, 30] < 0.00124 weeks (and 0 days)–28 weeks (and 6 days) 2634 (49.6%) 3198 (44.2%) 682 (43.8%) < 0.00129 weeks (and 0 days)–31 weeks (and 6 days) 2680 (50.4%) 4040 (55.8%) 874 (56.2%)Male 2908/5304 (54.8%) 3974/7233 (54.9%) 766 (49.9%) < 0.001Missing 10 5 3Singleton 3644 (68.6%) 5095 (70.4%) 1189 (76.5%) < 0.001Missing 0 1 1Small gestational age 545 (10.3%) 541 (7.5%) 277 (17.8%) < 0.001Missing 8 5 3Maternal chorioamnionitis 995 (23.5%) 1046 (21.0%) 168 (15.0%) < 0.001Missing 1085 2264 439Antenatal corticosteroids 5127 (96.9%) 5827 (81.7%) 1477 (95.8%) < 0.001Missing 23 108 14Outborn 299 (5.6%) 1484 (20.5%) 128 (8.2%) < 0.001Missing 0 6 2Delivered by caesarean 3082 (58.1%) 4260 (59.0%) 1119 (72.2%) < 0.001Missing 7 13 5Intensive resuscitation 1630 (30.9%) 2828 (39.6%) 466 (30.3%) < 0.001Missing 33 91 19Adjusted OR (vs no use)* 0.63 [0.54, 0.73] Reference – < 0.001Adjusted OR (vs other use)* 0.81 [0.66, 0.99] – Reference 0.04*Adjusted for gestational age, gender, small for gestational age, singleton, outborn status, delivery by caesarean and administration of corticosteroids.Chorioamnionitis was not included due to a large proportion of missing variablesDe Silva et al. Implementation Science  (2018) 13:8 Page 13 of 16heart rate [12], and reassuring results have been pub-lished from neurodevelopmental follow-up (includingintelligence quotient [IQ] measurement) of childrenfrom the original NP trials [62, 63].Strengths and limitations of the dataTo our knowledge, this is the first multicentre KT initia-tive of MgSO4 for fetal neuroprotection and of implemen-tation of national SOGC guidelines using a multifacetedstrategy. The CPN and MAG-CP datasets represent alarge population of women who presented with threatenedpreterm birth, and the sites that have contributed datarepresent most (78%) Canadian tertiary perinatal centres,of various sizes and from different geographic regions. KTactivities were well-documented prospectively and over asufficient period of time (i.e. 11 years) so that trends inMgSO4 use for fetal NP could be analysed according tocritical KT events. We used an interrupted time-seriesanalysis, which is a powerful quasi-experimental studydesign, to evaluate the effect of the KT intervention anddistinguish it from any observed effects in the absence ofintervention [32, 33]. Our prospective data collection in theCPN was detailed, included timing and dosage of MgSO4,pregnancy characteristics and maternal complications andfetal and neonatal outcomes. Also, we included data fromthe CNN, which allowed us to expand our analyses to non-MAG-CP sites and examine the impact (adjusted forconfounders) of MgSO4 for fetal NP on delivery roomintensive neonatal resuscitation, an outcome not availablein the CPN. As such, we believe our KT findings are gener-alisable to other clinicians who administer MgSO4 andmanage threatened very preterm birth, decision makersand researchers wishing to implement a national maternitycare clinical practice guideline or change practice.Among our limitations is the fact that site investigatorshad to report some (but not all) local KT activities, raisingthe possibility that some activities affecting MgSO4 usemay have been either over-reported or missed. Second,although our sample size of women was large overall,when examining effects within individual sites, or theassociation between individual components of the KTbundle compared with the overall bundle, we lackedpower [64]; we were unable to confirm which strategieswere responsible for the change in practice reported.Further, some aspects of the KT intervention were appliedat the same time across all centres (i.e. the SOGC guide-lines, the e-learning module and invitations to centralMAG-CP activities, such as newsletters and monthly tele-conferences); however, the application of other aspects ofKT were applied at different (non-random) times acrosssites (such as site visits and local rounds). It is also pos-sible that these KT strategies worked synergistically ratherthan the sum of effects by its individual components [65].Nevertheless, our segmented regression analysis shows anincrease in the optimal use during the KT interventionperiod. Moreover, the CNN data (from 31 sites) indicatedthat sites that participated in MAG-CP (vs those that didnot) had higher optimal use rates. Third, relating thematernal adverse effects and neonatal resuscitationoutcomes of MgSO4 in the same population of subjectswould have been optimal; however, we were unable todirectly link MAG-CP data with CNN.ConclusionsOptimal use of MgSO4 for fetal NP in Canada increasedsignificantly over 4 years with a multifaceted KT strategythat included education, engagement of health care pro-fessionals and identification of barriers and facilitators bythe local team. We have demonstrated that it is possibleto move from evidence-based national policy to imple-mentation. Our central support of the local KT teams is amodel worthy of consideration when planning implemen-tation of other clinical practice guidelines, whether local,regional or national. Specific to MgSO4 for fetal NP,future work should explore between-centre variability inpractice, the resolution of which may aid in achieving thetarget of 80% optimal use of MgSO4 for fetal NP. Weawait the results of pediatric motor and neurodevelop-mental outcomes associated with antenatal MgSO4 forfetal NP that are being tracked by the Canadian NeonatalFollow-Up Network at CPN and CNN sites [66].In general, future work should explore which compo-nents of a multifaceted strategy are particularly usefulfor implementing certain types of health interventions,such as drug interventions or surgical manoeuvres.Additional fileAdditional file 1: Figure S1. Between-centre variability of optimal,under and suboptimal uses of MgSO4 among the 11 participating centres.Figure S2. Trend of antenatal corticosteroid administration over timeamong women with underuse of MgSO4 as a proxy for non-precipitousdeliveries. Figure S3. Variability of MgSO4 use for fetal neuroprotectionamong MAG-CP centres (represented by the blue triangles) and non-MAG-CP centres (represented by the green circles). Table S1. MAG-CP (MAGnesiumsulphate for fetal neuroprotection to prevent Cerebral Palsy), CPN (CanadianPerinatal Network) and CNN (Canadian Neonatal Network) collaborativegroups. Table S2. Definitions of conditions and variables as used in theCanadian Perinatal Network (CPN). Table S3. Geographic regions of participatingcentres in the Canadian Perinatal Network (CPN). Table S4. Absolute utilisationrates of MgSO4 for fetal NP by study time period (from August 01/05 toSeptember 30/15). Table S5. Segmented regression analysis of the ninecentres that contributed data to both pre-MAG-CP and MAG-CP eras. TableS6. Sensitivity analyses of overall utilisation rates of MgSO4 using data fromthe nine centres that contributed data to both pre-MAG-CP and MAG-CPeras. Table S7. Determinants of engagement of participating sites in MAG-CP.Table S8. Components of engagement and relation to optimal use. Table S9.Antenatal MgSO4 use at GA 24–31+6 weeks by indication, from Jan 1/11 toSep 30/15. (DOCX 281 kb)AbbreviationsB&F: Barriers and Facilitators; CIHR: Canadian Institutes of Health Research;CNN: Canadian Neonatal Network; CP: Cerebral palsy; CPN: CanadianDe Silva et al. Implementation Science  (2018) 13:8 Page 14 of 16Perinatal Network; GEE: Generalised estimating equations; IUGR: Intrauterinegrowth restriction; KT: Knowledge translation; MAG-CP: MAGnesium sulphatefor fetal neuroprotection to prevent Cerebral Palsy; MgSO4: Magnesiumsulphate; NICU: Neonatal intensive care unit; NP: Neuroprotection;PPROM: Preterm pre-labour rupture of membranes; SOGC: Society ofObstetricians and Gynaecologists of CanadaAcknowledgementsWe acknowledge all members in the MAG-CP, CPN, and CNN CollaborativeGroups (Additional file 1: Table S1) for their part in collecting data.FundingThis project was funded by the Canadian Institutes for Health Research(MOP-115116). The funders played no role in the preparation of themanuscript or the decision to publish.Availability of data and materialsThe datasets analysed in this current study is available from thecorresponding author on reasonable request.Authors’ contributionsLAM, ARS and PvD conceived the project. LAM and DAD managed andcarried implementation of the project, with help from ARS and MAG-CPand CPN Collaborative Groups. All members in the MAG-CP, CPN andCNN Collaborative Groups (Additional file 1: Table S1) collected data for thestudy and the SteeringCommittee approved the analysis plan. DAD and TL analysed the data withadditional input from the data team. DAD, ARS and LAM wrote the first draftof the paper, with all other authors providing a critical review of subsequentdrafts. All authors approved the final manuscript.Ethics approval and consent to participateThe project was approved centrally by the Research Ethics Board at theUniversity of British Columbia (H05-70359 and H11-02214). As this wasapproved as a quality improvement project with no patient contact,patient consent was not deemed necessary.Consent for publicationNot applicable.Competing interestsAll authors declare that they have no competing interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Department of Obstetrics and Gynaecology, University of British Columbia,Vancouver, Canada. 2BC Children’s Hospital Research Institute, University ofBritish Columbia, Vancouver, Canada. 3Department of Paediatrics, Universityof British Columbia, Vancouver, Canada. 4Department of Women andChildren’s Health, St Thomas’ Hospital, 10th Floor, North Wing, WestminsterBridge Road, London SE1 7EH, UK. 5School of Life Course Sciences, Faculty ofLife Sciences and Medicine, King’s College London, London, UK.Received: 23 February 2017 Accepted: 26 December 2017References1. WHO recommendations on interventions to improve preterm birthoutcomes. Available at: http://www.who.int/iris/handle/10665/183037.Accessed 1 Feb 2017.2. Synnes A. Developmental Outcome. In: McDonald MG, Seshia M, editors.Avery’s neonatology: pathophysiology and management of the newborn.7th ed. Philadelphia: Wolters Kluwer; 2015. p. 1157–68.3. Chang E. Preterm birth and the role of neuroprotection. BMJ. 2015;350:g6661. https://doi.org/10.1136/bmj.g6661.4. Synnes AR, Anson S, Arkesteijn A, et al. School entry age outcomes forinfants with birth weight ≤ 800 grams. J Pediatr. 2010;157:994.e1. https://doi.org/10.1016/j.jpeds.2010.06.016.5. Mittendorf R, Dambrosia J, Pryde PG, et al. Association between the use ofantenatal magnesium sulfate in preterm labor and adverse health outcomes ininfants. Am J Obstet Gynecol 2002;186:1111-1118 doi:S0002937802000030 [pii].6. Crowther CA, Hiller JE, Doyle LW, et al. Effect of magnesium sulfate givenfor neuroprotection before preterm birth: a randomized controlled trial.JAMA. 2003;290:2669–76. https://doi.org/10.1001/jama.290.20.2669.7. Marret S, Marpeau L, Zupan-Simunek V, et al. Magnesium sulphate given beforevery-preterm birth to protect infant brain: the randomised controlled PREMAGtrial. BJOG. 2007;114:310–8. https://doi.org/10.1111/j.1471-0528.2006.01162.x.8. Rouse DJ, Hirtz DG, Thom E, et al. A randomized, controlled trial ofmagnesium sulfate for the prevention of cerebral palsy. N Engl J Med. 2008;359:895–905. https://doi.org/10.1056/NEJMoa0801187.9. Conde-Agudelo A. Antenatal magnesium sulfate for the prevention ofcerebral palsy in preterm infants less than 34 weeks’ gestation: a systematicreview and metaanalysis. Am J Obstet Gynecol. 2009;200:595–609. https://doi.org/10.1016/j.ajog.2009.04.005.10. Costantine MM, Weiner SJ, Eunice Kennedy Shriver National Institute ofChild Health and Human Development Maternal-Fetal Medicine UnitsNetwork. Effects of antenatal exposure to magnesium sulfate onneuroprotection and mortality in preterm infants: a meta-analysis. ObstetGynecol 2009;114:354-364 doi:https://doi.org/10.1097/AOG.0b013e3181ae98c2 [doi].11. Doyle LW, Crowther CA, Middleton P, et al. Magnesium sulphate for womenat risk of preterm birth for neuroprotection of the fetus. Cochrane DatabaseSyst Rev. 2009:CD004661. https://doi.org/10.1002/14651858.CD004661.pub3.12. Nensi A, De Silva DA, von Dadelszen P, et al. Effect of magnesium sulphateon fetal heart rate parameters: a systematic review. J Obstet Gynaecol Can2014;36:1055-1064 doi:S1701–2163(15)30382–0 [pii].13. Drassinower D, Friedman AM, Levin H, et al. Does magnesium exposureaffect neonatal resuscitation? Am J Obstet Gynecol. 2015;213:424.e5. https://doi.org/10.1016/j.ajog.2015.05.052.14. Jacquemyn Y, Zecic A, Van Laere D, et al. The use of intravenousmagnesium in non-preeclamptic pregnant women: fetal/neonatalneuroprotection. Arch Gynecol Obstet. 2015;291:969–75. https://doi.org/10.1007/s00404-014-3581-1 [doi].15. De Silva DA, Sawchuck D, von Dadelszen P, et al. Magnesium sulphate foreclampsia and fetal neuroprotection: a comparative analysis of protocolsacross Canadian tertiary perinatal centres. J Obstet Gynaecol Can 2015;37:975-987 doi:S1701–2163(16)30047–0 [pii].16. Teela KC, De Silva DA, Chapman K, et al. Magnesium sulphate for fetalneuroprotection: benefits and challenges of a systematic knowledgetranslation project in Canada. BMC Pregnancy Childbirth. 2015;15:8. https://doi.org/10.1186/s12884-015-0785-8.17. Chollat C, Le Doussal L, de la Villeon G, et al. Antenatal magnesium sulphateadministration for fetal neuroprotection: a French national survey. BMCPregnancy Childbirth. 2017;17:304.18. Magee LA, Hall D, van der Merwe JL, et al. Chapter 8: fluids, drugs andtransfusion. In: Magee LA, von Dadelszen P, Stones W, et al., editors. TheFIGO textbook of hypertension in pregnancy: an evidence-based guide tomonitoring, prevention and management. Croyden: GLOWM; 2016.19. Rogers EM. Diffusion of innovations: Simon and Schuster 2003.20. Nilsen P. Making sense of implementation theories, models and frameworks.Implement Sci. 2015;10:53. https://doi.org/10.1186/s13012-015-0242-0.21. Wensing M, Bosch M, Grol R. Developing and selecting interventions fortranslating knowledge to action. CMAJ. 2010;182:85. https://doi.org/10.1503/cmaj.081233.22. Damarell RA, Tieman JJ. How do clinicians learn about knowledgetranslation? An investigation of current web-based learning opportunities.JMIR Med Educ. 2017;3:e12.23. Jamtvedt G, Young JM, Kristoffersen DT, et al. Does telling people what theyhave been doing change what they do? A systematic review of the effectsof audit and feedback. Quality & safety in health care. 2006;15(6):433.https://doi.org/10.1136/qshc.2006.018549.24. Hysong SJ. Meta-analysis: audit and feedback features impact effectivenesson care quality. Med Care. 2009;47:356–63. https://doi.org/10.1097/MLR.0b013e3181893f6b.25. Johnston G, Crombie IK, Davies H, et al. Reviewing audit: barriers andfacilitating factors for effective clinical audit. Qual Health Care. 2000;9:23–36.26. Magee L, Sawchuck D, Synnes A, et al. SOGC Clinical Practice Guideline.Magnesium sulphate for fetal neuroprotection. J Obstet Gynaecol Can 2011;33:516-529 doi:S1701–2163(16)34886–1 [pii].De Silva et al. Implementation Science  (2018) 13:8 Page 15 of 1627. Pinnock H, Barwick M, Carpenter CC, et al. Standards for ReportingImplementation Studies (StaRI) statement. BMJ. 2017;356:i6795.28. Cane J, O’Connor D, Michie S. Validation of the theoretical domainsframework for use in behaviour change and implementation research.Implement Sci. 2012;7:37. https://doi.org/10.1186/1748-5908-7-37.29. The Canadian Perinatal Network. Available at: www.cpn-rpc.org., 2016.30. Carroll C, Patterson M, Wood S, et al. A conceptual framework forimplementation fidelity. Implement Sci. 2007;2:40. https://doi.org/10.1186/1748-5908-2-40.31. Magee LA, von Dadelszen P, Allen VM, et al. The Canadian PerinatalNetwork: a national network focused on threatened preterm birth at 22 to28 weeks’ gestation. J Obstet Gynaecol Can 2011;33:111-120 doi:S1701–2163(16)34795–8 [pii].32. Taljaard M, McKenzie JE, Ramsay CR, et al. The use of segmentedregression in analysing interrupted time series studies: an example inpre-hospital ambulance care. Implement Sci. 2014;9:77. https://doi.org/10.1186/1748-5908-9-77.33. Penfold RB, Zhang F. Use of interrupted time series analysis in evaluatinghealth care quality improvements. Acad Pediatr. 2013;13:38.34. The R Project for statistical computing. 2017. Available at: https://www.r-project.org.35. Weisz DE, Shivananda S, Asztalos E, et al. Intrapartum magnesium sulfate andneed for intensive delivery room resuscitation. Arch Dis Child Fetal NeonatalEd. 2015;100:59. https://doi.org/10.1136/archdischild-2013-305884 [doi].36. Knowledge Translation in Health Care. Moving from evidence to practice.West Sussex: Wiley-Blackwell; 2013.37. Chaillet N, Dube E, Dugas M, et al. Evidence-based strategies forimplementing guidelines in obstetrics: a systematic review. Obstet Gynecol2006;108:1234-1245 doi:108/5/1234 [pii].38. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: asystematic review of rigorous evaluations. Lancet 1993;342:1317-1322 doi:0140–6736(93)92244-N [pii].39. Tamblyn R, Battista R. Changing clinical practice: which interventions work?J Contin Educ Heal Prof. 1993;13:273–88. https://doi.org/10.1002/chp.4750130403.40. Thanh NX, Jacobs P, Wanke MI, et al. Outcomes of the introduction of theMOREOB continuing education program in Alberta. J Obstet Gynaecol Can.2010;32:749–55. doi:S1701-2163(16)34615-1 [pii]41. McGuire W, Fowlie PW. Bridging the gaps: getting evidence into practice.CMAJ. 2009;181:457–8. https://doi.org/10.1503/cmaj.091243.42. Bain E, Bubner T, Ashwood P, et al. Barriers and enablers to implementingantenatal magnesium sulphate for fetal neuroprotection guidelines: a studyusing the theoretical domains framework. BMC Pregnancy Childbirth. 2015;15:9. https://doi.org/10.1186/s12884-015-0618-9.43. Wensing M, Grol R. Single and combined strategies for implementingchanges in primary care: a literature review. Int J Qual Health Care.1994;6:115–32.44. Wensing M, van der Weijden T, Grol R. Implementing guidelines andinnovations in general practice: which interventions are effective? Br J GenPract. 1998;48:991–7.45. Oxman AD, Thomson MA, Davis DA, et al. No magic bullets: a systematicreview of 102 trials of interventions to improve professional practice. CMAJ.1995;153:1423–31.46. Squires JE, Sullivan K, Eccles MP, et al. Are multifaceted interventions moreeffective than single-component interventions in changing health-careprofessionals’ behaviours? An overview of systematic reviews. ImplementSci. 2014;9:152. https://doi.org/10.1186/s13012-014-0152-6.47. American College of Obstetricians and Gynecologists Committee onObstetric Practice, Society for Maternal-Fetal Medicine. Committee opinionno. 455: magnesium sulfate before anticipated preterm birth forneuroprotection. Obstet Gynecol. 2010;115:669–71. https://doi.org/10.1097/AOG.0b013e3181d4ffa5.48. The Antenatal Magnesium Sulphate for Neuroprotection GuidelineDevelopment Panel. Antenatal magnesium sulphate prior to preterm birthfor neuroprotection of the fetus, infant and child: national clinical practiceguidelines. Adelaide: The University of Adelaide; 2010. Available at: https://www.adelaide.edu.au/arch/antenatalMagnesiumSulphateGuidlines.pdf.Accessed 1 Feb 2017.49. Royal College of Obstetricians & Gynaecologists. Magnesium wsulphate toprevent cerebral palsy following preterm birth. Available at: https://www.rcog.org.uk/globalassets/documents/guidelines/scientific-impact-papers/sip_29.pdf. Accessed 1 Feb 2017.50. Gibbins KJ, Browning KR, Lopes VV, et al. Evaluation of the clinical use ofmagnesium sulfate for cerebral palsy prevention. Obstet Gynecol. 2013;121:235–40. https://doi.org/10.1097/AOG.0b013e31827c5cf8.51. Bouet PE, Brun S, Madar H, et al. Implementation of an antenatalmagnesium sulfate protocol for fetal neuroprotection in preterm infants. SciRep. 2015;5 https://doi.org/10.1038/srep14732, https://doi.org/10.1038/srep14732.52. Tan YH, Groom KMA. Prospective audit of the adherence to a newmagnesium sulphate guideline for the neuroprotection of infants born lessthan 30 weeks’ gestation. Aust N Z J Obstet Gynaecol. 2015;55:90–3. https://doi.org/10.1111/ajo.12271 [doi].53. Crowther CA, Middleton PF, Bain E, et al. Working to improve survival and healthfor babies born very preterm: the WISH project protocol. BMC PregnancyChildbirth. 2013;13:239. https://doi.org/10.1186/1471-2393-13-239 [doi].54. Siwicki K, Bain E, Bubner T, et al. Nonreceipt of antenatal magnesiumsulphate for fetal neuroprotection at the Women’s and Children’s Hospital,Adelaide 2010–2013. Aust N Z J Obstet Gynaecol. 2015;55:233–8. https://doi.org/10.1111/ajo.12334 [doi].55. Oddie S, Tuffnell DJ, McGuire W. Antenatal magnesium sulfate: neuro-protection for preterm infants. Arch Dis Child Fetal Neonatal Ed. 2015;100:553. https://doi.org/10.1136/archdischild-2014-307655 [doi].56. Pang J. Adherence to uptake of magnesium sulphate for neuroprotection inpreterm births. N Z Med J. 2017;130:52–3.57. Wolf HT, Huusom L, Weber T, et al. Use of magnesium sulfate before32 weeks of gestation: a European population-based cohort study. BMJOpen. 2017;7:e013952. https://doi.org/10.1136/bmjopen-2016-013952.58. Lipsitz PJ, English IC. Hypermagnesemia in the newborn infant. Pediatrics.1967;40:856–62.59. Lloreda-Garcia JM, Lorente-Nicolas A, Bermejo-Costa F, et al. Need forresuscitation in preterm neonates less than 32 weeks treated with antenatalmagnesium sulphate for neuroprotection. Rev Chil Pediatr. 2016;87:261–7.https://doi.org/10.1016/j.rchipe.2015.11.006 [doi].60. Johnson LH, Mapp DC, Rouse DJ, et al. Association of cord bloodmagnesium concentration and neonatal resuscitation. J Pediatr. 2012;160:577.e1. https://doi.org/10.1016/j.jpeds.2011.09.016.61. De Jesus LC, Sood BG, Shankaran S, et al. Antenatal magnesium sulfateexposure and acute cardiorespiratory events in preterm infants. Am JObstet Gynecol. 2015;212:94.e7. https://doi.org/10.1016/j.ajog.2014.07.023.62. Chollat C, Enser M, Houivet E, et al. School-age outcomes following arandomized controlled trial of magnesium sulfate for neuroprotection ofpreterm infants. J Pediatr. 2014;165:400.e3. https://doi.org/10.1016/j.jpeds.2014.04.007.63. Doyle LW, Anderson PJ, Haslam R, et al. School-age outcomes of very preterminfants after antenatal treatment with magnesium sulfate vs placebo. JAMA.2014;312:1105–13. https://doi.org/10.1001/jama.2014.11189 [doi].64. Stevens BJ, Yamada J, Promislow S, et al. Implementation ofmultidimensional knowledge translation strategies to improve proceduralpain in hospitalized children. Implement Sci. 2014;9:1. https://doi.org/10.1186/s13012-014-0120-1 [doi].65. Moore GF, Audrey S, Barker M, et al. Process evaluation of complexinterventions: Medical Research Council guidance. BMJ. 2015;35066. Synnes A, Luu TM, Moddemann D, et al. Determinants of developmentaloutcomes in a very preterm Canadian cohort. Arch Dis Child Fetal NeonatalEd. 2016; doi:fetalneonatal-2016-311228 [pii]•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:De Silva et al. Implementation Science  (2018) 13:8 Page 16 of 16


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items