WIXLIB

BMJ Global HealthSearch strategiesSearches were conducted between November andDecember 2020. Databases searched were: PubMed/MEDLINE; Academic Search Complete (EBSCO);EMBASE; CINAHL (EBSCO); Web of Science; Scopus;Nursing and Allied Health (ProQuest); Maternity andInfant Care (Ovid). We conducted an online hand searchby checking reference lists of identified articles. With thesupport of librarians, a search strategy was created usingpredefined terms and synonyms for the intervention,population and setting (online supplemental file 2).Study selectionTitles and abstracts of all search results were importedinto Endnote and duplicates removed. The online reviewsoftware Covidence was used to screen titles, abstractsand full texts.28 Two authors independently reviewedeach title and abstract against the eligibility criteria, withpotentially relevant articles included for full- text review.Full texts were recovered and independently assessedfor eligibility by two reviewers. Disagreements at anystage were resolved by discussion or by involving a thirdreviewer. Where more than one paper reported the samestudy, the papers were collated to ensure the primarystudy was the unit of interest.We used the Mixed Method Appraisal Tool (MMAT)for assessing the quality of studies.29 MMAT is a qualityassessment instrument for quantitative, qualitative andmixed- methods studies in a single tool.30Homer CSE, et al. BMJ Global Health 2022;7:e006872. doi:10.1136/bmjgh-2021-006872Patient and public involvementPatients and the public were not directly involved in thisreview.AnalysisThe findings of each study were imported into NVivoV.12. Coding process was undertaken with a focus onour two research objectives: factors affecting an enablingenvironment and the availability and effectiveness ofsupport systems for midwives.We initially took an inductive approach. We identifiedand extracted categories and concepts that emergedfrom the data and developed codes highlighting the keyissues. As we proceeded through the coding, additionalideas emerged which were organised as concepts.31 Thenext stage was to undertake a deductive process using ourtwo objectives to identify other concepts relating to theseareas specifically. A mind map was developed in NVivoto connect the concepts with our objectives and builda comprehensive picture of the issues.32 Our objectivesdeveloped into the two key themes and we identified aseries of subthemes.We summarised the findings by type of publication,method, context of the study, type of intervention andlevel of fragility. As studies varied in their method,context and outcomes used, it was not possible to pooldata for a formal meta- analysis.33 We therefore undertook a narrative synthesis.31 Narrative synthesis has beenused in previous systematic reviews where the quantitative and qualitative results are integrated to inform thefindings.34 35 We did not undertake any assessment ofmeta- bias(es), such as publication bias across studies orselective reporting within studies.RESULTSIn total, 2500 citations were identified and imported intoEndnote and Covidence (table 1).18 36 After irrelevantand duplicate citations were removed, 461 potentiallyeligible papers were identified (figure 1). We were ableto screen the full text of 422 papers, identifying 39 papersfor assessment. During a data extraction and qualityassessment phase, a further 15 papers were excluded,as there was inadequate or vague information about thestudy population or interventions. For example, it wasunclear whether midwives were included in the study ortheir contribution could not be identified. Ultimately,24 papers were included in the review. Included articleswere summarised and categorised according to their relevance to the objectives of the study (table 2).Of the 24 studies, 14 were qualitative, 7 mixedmethods and 3 quantitative. Most studies were descriptive (n 16). Of the countries included, seven are low- income countries and three are lower- middle- incomecountries.21 Three countries are categorised as havinghighest fragility, two as very high fragility, two as highfragility and three as fragile settings. These countriesincluded Afghanistan,37–43 Bangladesh,44–48 Nigeria,49–523BMJ Glob Health: first published as 10.1136/bmjgh-2021-006872 on 20 January 2022. Downloaded from http://gh.bmj.com/ on January 21, 2022 by guest. Protected by copyright.discussions, individual interviews, observation, diaries,document analysis, open- ended survey questions) andqualitative methods for data analysis (eg, thematicanalysis, framework analysis, grounded theory). Wealso included randomised trials, non- randomised trials,controlled before- and- after studies, interrupted timeseries and repeated measures26 and mixed- methodsstudies, provided they were primary studies and relevant to the review question.The population of interest was midwives/nurse- midwives,according to the scope of practice defined by the International Confederation of Midwives,27 as well as the women,girls and children in their care. Any practices or interventions that targeted health workers working outside thescope of midwifery practice were excluded. Interventionsthat focused on nurses, obstetricians and gynaecologists,paediatricians or community health workers were excluded.Eligible studies were categorised according to the four categories: highest fragility, very high fragility, high fragility orfragile settings.We excluded case reports or case series, letters, editorials,commentaries, reviews, study protocols, conference abstractsor other article types that did not provide primary data and/or were not peer- reviewed. Specific cost- effectiveness studieswere also excluded but studies that included costing data aspart of implementation and feasibility were included.

BMJ Global HealthData sourceResultsDuplicateremovedUnrelated articlesremovedReviewed abstractsand titlesCINAHL (EBSCO)Pubmed/Medline237236185236113766858Web of Science2682478663Nursing and Allied Academic Search Complete (EBSCO)74469822812287763737NA21267792461Maternity & Infant Care (OVID)Hand searchedTotal232477 ( 23 handsearch)NA, not available.Democratic Republic of Congo (DRC),53 54 South Sudanand Sudan,55 56 Ethiopia, Pakistan, Uganda and Liberia.Afghanistan (n 7), Bangladesh (n 5), Nigeria (n 4),DRC (n 2) made up most of the studies (table 2). TheFigure 14definition and scope of practice for midwives were oftenunclear for some countries, such as Nigeria, SouthSudan, DRC and Ethiopia. However, studies in Afghanistan, Bangladesh, Pakistan, Sudan and Liberia specifiedPRISMA flow chart showing search results, screening and included papers.Homer CSE, et al. BMJ Global Health 2022;7:e006872. doi:10.1136/bmjgh-2021-006872BMJ Glob Health: first published as 10.1136/bmjgh-2021-006872 on 20 January 2022. Downloaded from http://gh.bmj.com/ on January 21, 2022 by guest. Protected by copyright.Table 1 Results of database search

Target groupMidwives anddoctorsCompareMidwiferyperformancegraduates andof midwiferystudentsgraduates whowere selectedthroughcommunitymobilisationwith thosewho admittedthrough a nationalexamination.Assessmentof emergencyobstetric andnewborn caretraining/servicesAssessment ofMidwives, doctorsmidwifery services and care assistantsto explore cultureof care.Assess midwifery Midwives, doctorsservices toand senior officialsexplore underlyingissues for staffbehavioursInterventionMixed method/retrospectivesurvey.MW studentrecords of the178 trainees werereviewed anddata extracted fordemographics,selection method,performancein knowledgeand skills, anddeployment statusand location.Quantitative/cross- sectionalstudy142 midwives and82 doctorsQualitative/ethnographicstudy23 Interviews wereconducted withsenior doctorsand midwives,resident doctors,newly qualifiedmidwives andcare assistants (2senior midwives8 midwives with 6months–10 23 Interviews wereconducted withsenior doctorsand midwives,resident doctors,newly qualifiedmidwives and careassistantsMethodologyContinued96% of midwifery graduates selected by communities wereemployed, compared with 74% chosen by the Institute of HealthSciences (IHS) and 82% by the National University EntranceExamination (NUEE). 63% of community- selected graduates wereworking in rural locations, compared with 43% recruited by IHSand 9% by the NUEE. While fewer midwifery graduates selectedby communities had completed high school and their academicperformance was slightly lower during training, there was nodifference in their pass rates and acquisition of practical skills.Over 90% of facilities had essential equipment for newbornresuscitation, including a mucus extractor, bag and mask. Morethan 80% of providers had been trained on newborn resuscitation,but midwives were more likely than doctors to receive such trainingas part of preservice education (59% and 35%, respectively,p 0.001). Training was associated with greater knowledge(p 0.001) and clinical skills (p 0.05) in a multivariable model thatadjusted for facility type, provider type and years of experienceoffering emergency obstetric and newborn care services.Most staff members were simply endeavouring to survive in atough working environment where the lack of a shift systeminevitably resulted in staff exhaustion, poor performance and theconstant risk of mistakes leading to censure by management.Doctors and midwives concurred that they did not provide careas they had been taught and blamed the workload, lack of ashift system, insufficient supplies and inadequate support frommanagement.Although healthcare providers generally look the same in hospitalsacross the globe, this study reveals they have different ways ofunderstanding the world. They may work in a ‘caring’ professionbut perceive the purpose of work from innate social norms andvalues, which, for example, can result in guarding their skills ratherthan sharing them. A sense of responsibility to strangers mightclash with kinship obligations, the institutional culture, and thewider social environment. Furthermore, public health institutionsmay not only provide care, they may also be lucrative sources ofincome for political elites.Results/outcomesBMJ Glob Health: first published as 10.1136/bmjgh-2021-006872 on 20 January 2022. Downloaded from http://gh.bmj.com/ on January 21, 2022 by guest. Protected by copyright.Homer CSE, et al. BMJ Global Health 2022;7:e006872. doi:10.1136/bmjgh-2021-006872Highest Fragility Midwifery trainingin post- conflictAfghanistan: tensionsbetween educationalstandards and ruralcommunity needsAfghanistanMansoor, 2012404Highest Fragility Assessing thecapacity for newbornresuscitation andfactors associatedwith providers'knowledge and skills:a cross- sectionalstudy in AfghanistanAfghanistanKim,2013393Highest Fragility Villains or victims?An ethnography ofAfghan maternity staffand the challenge ofhigh- quality eclassificationHighest Fragility Parallel worlds: Anethnography of care inan Afghan nFirst author, yearCharacteristics of included studies#Table 2BMJ Global Health5

Homer CSE, et al. BMJ Global Health 2022;7:e006872. doi:10.1136/bmjgh-2021-006872Zainullah, 201443Bogren,20184578TitleInterventionFragile SettingBangladeshOpportunities,challenges andstrategies whenbuilding a midwiferyprofession. Findingsfrom a qualitativestudy in Bangladeshand NepalHighest Fragility Establishing midwiferyin low- resourcesettings: Guidancefrom a mixed- methodsevaluation of theAfghanistan midwiferyeducation programmeHighest Fragility Factors influencing theretention of midwivesin the public sectorin Afghanistan: Aqualitative assessmentof midwives in eightprovincesWomenTarget groupAssess strategiesfor establishmentof midwiferyEducationMidwivesEvaluation of costs Midwivesand graduateperformanceoutcomes ofthe two typesof preservicemidwiferyeducationprogrammesEvaluation ofMidwives and otherfactors associated stakeholderswith midwifery(MW) deploymentHighest Fragility ‘The midwife helpedTo exploreexperiences ofme. Otherwise Imidwifery carecould have died’:women’s experienceof professionalmidwifery services inrural Afghanistan—aqualitative study in theprovinces Kunar tan6Wood,201342Thommesen, 202041 Afghanistan5CountriesFirst author, yearContinuedQualitativeQuantitativeanalysis ofmidwiferygraduates’performance andcost analysisto estimate theresource requiredto educate amidwife. midwiferyschool graduates(n 138)QualitativeStakeholdersn 33, MWstudents n 35,MW currentlyemployed n 17;MWs formerlyemployed n 16,CHWs n 16Qualitative/exploratory study39 womenparticipated—25who had givenbirth during thelast 6 months,11 mothers- in- law and threecommunitymidwives in theprovinces of Kunarand LaghmanMethodologyContinuedGlobal and national standards brought together midwiferyprofessionals and stakeholders, and helped in the establishmentof midwifery associations. The challenges were nationalcommitments without a full set of supporting policy documents,lack of professional recognition, and competing views, interestsand priorities.Graduates (n 101) achieved an overall mean competency scoreof 63.2% (59.9%–66.6%) on the clinical competency assessmentcompared with 57.3% (49.9%–64.7%) for Institute of HealthSciences graduates (n¼37). Reproductive health activitiesaccounted for 76% of midwives' time over an average of 3months. Approximately 1% of childbirths required referral orresulted in maternal death. On the basis of known costs for theprogrammes, the estimated cost of graduating a class with 25students averaged US 298 939, or US 10 784 per graduate.Several factors affect a midwife throughout her career in thepublic sector, including her selection as a trainee, the trainingitself, deployment to her preassigned post, and working in clinics.Overall, appropriate selection is the key to ensuring deploymentand retention later on in a midwife’s career. Other factors thataffect retention of midwives include civil security concerns in ruralareas, support of family and community, salary levels, professionaldevelopment opportunities and workplace support, and inefficienthuman resources planning in the public sector.Many of the women greatly valued the trained midwives’ life- savingexperience, skills and care, and the latter were important reasonsfor choosing to give birth in a clinic. Women further appreciatedmidwives’ promotion of immediate skin- to- skin contact andbreastfeeding. However, some women experienced rudeness,discrimination and negligence on the part of the midwives.Moreover, relatives’ disapproval, shame and problems withtransport and security were important obstacles to women givingbirth in the clinics.Results/outcomesBMJ Glob Health: first published as 10.1136/bmjgh-2021-006872 on 20 January 2022. Downloaded from http://gh.bmj.com/ on January 21, 2022 by guest. Protected by copyright.6#Table 2BMJ Global Health

Homer CSE, et al. BMJ Global Health 2022;7:e006872. angladeshBangladeshBogren, 2018Bogren, 201847Bogren, 201548Zaman, 202044101112Countries9First author, yearContinuedFragile SettingFragile SettingFragile SettingFragile SettingFragileclassificationInterventionTarget groupMethodologyExperiences of a newcadre of midwives inBangladesh: findingsfrom a mixed methodstudyTowards a midwiferyprofession inBangladesh—asystems approach fora complex worldDevelopment of acontext- specificaccreditationassessment toolfor affirming qualitymidwifery education inBangladeshMidwivesAssessexperiencesof Midwiferystudents ontheir midwiferyeducation andtheir first clinicalpost as qualifiedmidwifeMidwivesAssess howMidwivesstakeholderspromote midwiferyprofessionAssess thefeasibility ofdevelopment of acontext- specificaccreditationassessment toolfor midwiferyeducationMixed methodInterviewing329 midwivesand conductingsix focus groupdiscussions with43 midwivesand midwiferystudents.Qualitative/explorative study.Governmentn 4; academian 3; Professionalassociation n 3;donors n 9; NGOsn 6Mixed methods(using quantitativeand descriptivequestionnaire).123 nursingeducatorsteaching the 3years diplomamidwiferyeducationprogramme.What preventsAssess midwifery Midwifery students Qualitativemidwifery quality care services in relation67 midwiferyin Bangladesh? Ato quality of carestudentsfocus group enquirywith midwiferystudentsTitleContinuedMost of the midwives were satisfied with different dimensionsof their education programme, with the exception of the levelof exposure they had to the rural communities during theirprogramme. Out of 329 midwives, 50% received tuition feewaivers, while 46% received funding for educational materialsand 40% received free accommodation. The satisfaction withthe various aspects of the current posting was high and nearly allmidwives reported a desire to work in the public sector in the longrun. However, a significant proportion of the midwives expressedconcerns with equipment, accommodation, transport and prospectof transfers.Collaboration between organisations was valued, as more couldbe achieved compared with what an individual organisation coulddo. Significant results of this were that two midwifery curricula andfaculty developments had been produced. Although collaborationwas mostly seen as something positive to move the systemforward, the approach to reach the set goal varied with differentinterests, priorities and concerns, both on individual organisationallevel and at system level. Frequent struggles between individualphilosophies and organisational mandates were seen as competinginterests for advancing the national priorities. It would appear thatnewcomers with innovative ideas were denied access on the sameterms as other actors.Provides insight into the development of a context- specificaccreditation assessment tool for Bangladesh. Importantcomponents to be included in this accreditation tool are presentedunder the following categories and domains: ‘organisation andadministration’, ‘midwifery faculty’, ‘student body’, ‘curriculumcontent’, ‘resources, facilities and services’ and ‘assessmentstrategies’. The identified components were a prerequisite toensure that midwifery students achieve the intended learningoutcomes of the midwifery curriculum, and hence contribute toa strong midwifery workforce. The components further ensurewell- prepared teachers and a standardised curriculum supportedat policy level to enable effective deployment of professionalmidwives in the existing health system.Professional barriers include heavy workloads with a shortage ofstaff who were not utilised to their full capacity within the healthsystem. The reason for this was a lack of recognition in themedical hierarchy, leaving midwives with low levels of autonomy.Economical barriers were reflected by lack of supplies and hospitalbeds, midwives earning only low and/or irregular sala

The importance of effective midwifery services is evident, yet there is limited knowledge about the inter-ventions, support systems and enabling environments that can enhance quality midwifery care in HFS. A supplement published with the 2021 State of the World's Midwifery Report presented three case studies from Bangladesh,