a systematic review of the effectiveness of interventions to

A systematic review of the effectiveness of interventions to increase the early initiation of antenatal care in socially disadvantaged and

vulnerable women

Final Report

Laura Oakley, Ron Gray, Jennifer J Kurinczuk, Peter Brocklehurst, Jennifer Hollowell

National Perinatal Epidemiology Unit, University of Oxford

October 2009

A systematic review of the effectiveness of interventions to increase the early initiation of antenatal care in socially disadvantaged and vulnerable women

ContentsExecutive summary .....................................................................................1

1 Background ...........................................................................................4

1.1 Aims of the review ..........................................................................6

2 Definitionsandscopeofthereview ...........................................................6

2.1 Operationaldefinitionofcomprehensiveantenatalcare ........................6

2.2 ‘Early’ initiation of antenatal care ......................................................6

2.3 Typesofintervention .......................................................................7

2.4 NHS relevance ................................................................................7

2.5 Disadvantagedandvulnerablegroups................................................7

3 Methods ................................................................................................8

3.1 Inclusion Criteria ............................................................................8

3.1.1 Study design ......................................................................8

3.1.2 Population ..........................................................................8

3.1.3 Intervention .......................................................................8

3.1.4 Comparatorgroup ...............................................................8

3.1.5 Outcome measure ...............................................................8

3.1.6 Language ...........................................................................9

3.1.7 Timeperiod ........................................................................9

3.1.8 Geographicalareas ..............................................................9

3.1.9 Typesofpublication .............................................................9

3.2 Exclusions .....................................................................................9

3.3 Methodsforidentificationofstudies ..................................................9

3.3.1 Overview of strategy to identify relevant studies ......................9

3.3.2 Bibliographicdatabases ......................................................10

3.3.3 Other online searchable resources .......................................10

3.3.4 Itemsidentifiedinscopingexerciseandantenatalcarereview .............................................................................11

3.3.5 Reference lists and citations ................................................11

3.4 Review methods ...........................................................................11

3.4.1 Screening .........................................................................11

3.4.2 Quality assessment ............................................................13

3.4.3 Data extraction .................................................................14

3.4.4 Assessment of effectiveness ................................................14

4 Results ................................................................................................14

4.1 Overview of included studies ..........................................................16

4.1.1 Countries .........................................................................16

4.1.2 Yearofpublication/study ....................................................16

4.1.3 Study design ....................................................................17


4.1.4 Outcome measure .............................................................17

4.1.5 Quality .............................................................................18

4.1.6 Assessment of whether timing of initiation of antenatal care was an outcome measure ............................................19

4.2 Interventions studied ....................................................................19

4.2.1 Interventionrecipients/targetpopulations .............................19

4.2.2 Intervention content ..........................................................19

4.3 Effectiveness ................................................................................24

4.3.1 Outreach or other community-based interventions .................24

4.3.2 Alternative models of clinic-based antenatal care ...................25

5 Discussion ...........................................................................................29

5.1 Principalfindings ..........................................................................29

5.2 Strengths and limitations of this systematic review............................30

5.3 Findingsinrelationtootherpublishedevidence ................................31

5.4 Implicationsandrecommendations .................................................32

5.5 Conclusion ...................................................................................33

Acknowledgement .....................................................................................33

References ...............................................................................................34

Annex A: Medline search strategy ................................................................40

Annex B: Named intervention searches ........................................................42

Annex C: Characteristics and results of included studies .................................43


TablesTable 1. Exclusion criteria ...........................................................................12

Table 2. Reasons for exclusion .....................................................................16

Table3.Majorflawsidentifiedinincludedstudies ...........................................18

Table 4. Effectiveness of outreach or other community-based interventions ............................................................................................26

Table 5. Effectiveness of interventions involving alternative models of clinic-based antenatal care .........................................................................28

FiguresFigure1.Barrierstoequitablehealthcareforracialandethnicgroups .................5

Figure2.Screeningprocess ........................................................................15

Figure3.Yearofpublicationofincludedstudies .............................................17

A systematic review of the effectiveness of interventions to increase the early initiation of antenatal care in socially disadvantaged and vulnerable women 1

Executive summaryThesystematicreviewdescribedinthisreportispartofaprogrammeofwork,commissionedbytheDepartmentofHealth,tostrengthentheevidencebaseoninterventionstoreduceinfantmortality,withaparticularfocusonreducinginequalitiesininfant mortality.

Aim

Thepurposeofthisreviewwastosystematicallyidentifyandevaluatetheevidencerelatingtotheeffectivenessofinterventions,relevantinthecontextoftheNationalHealthService(NHS),whichaimtoincreasetheearlyinitiationofcomprehensiveantenatalcarein socially disadvantaged and vulnerable women.

Methods

Searches

Wesearchedthemajorbibliographicdatabasesusingatwostagestrategy:weinitiallyranacomprehensive‘generic’searchandthenranfurthersearchesincorporatingtextsearchtermsrelatingtointerventionsidentifiedintheinitialsearches.Wealsosearchedotheronlinelibrariesandresources(e.g.CochraneLibrary,NationalGuidelinesClearingHouse)forrelevantsecondaryreports.Thereferencesandcitationsofincludedstudiesandrelevantsecondaryreportswerechecked.

Inclusion criteria

Studies which met the following “PICO” criteria were eligible for inclusion:

Population

Interventionevaluatedinarelevantdisadvantagedorvulnerablepopulation.•

PopulationrecruitedinanOECDcountry(excludingTurkeyandMexico).•

Intervention

Wedidnotplaceanyrestrictiononthetypeofintervention.Werequiredonlythat•studiesreportedthetimingofinitiationofantenatalcareasanoutcomemeasure.

Comparator

Studyincludedacontrol/comparatorgroup(s)whichdidnotreceiveorhaveaccessto•the intervention.

Interventionandcomparatorgroupwereselectedusingthesameand/orsimilar•samplingframesandbothgroupsdrawnfrombroadlysimilarpopulations.

Outcome

Theproportionofwomeninitiatingcomprehensiveantenatalcarebyagivenweek/•month(<=20weeksorbeforethefifthmonthofgestation).

Studiesrelatingsolelytotheprovisionorextensionofhealthinsurancecoverage,alongwith studies relating to models of insurance coverage or reimbursement were excluded. Wealsoexcludedstudiesprimarilyaddressingbarrierstoantenatalcareaccessthatrelatedtostructuralorfinancialaspectsofthelocalhealthcaresystemnotconsideredtoapplyinapredominantlygovernment-fundeduniversalhealthcaresystemsuchastheNHS.

A systematic review of the effectiveness of interventions to increase the early initiation of antenatal care in socially disadvantaged and vulnerable women2

Quality assessment

TworeviewersindependentlyappliedtheGATEchecklisttoassesstheinternalvalidityofeachstudy,focusingonthevalidityoftheestimatedeffectonthetimingofinitiationofantenatalcare.Usingthischecklist,theinternalvalidityofeachincludedstudywasratedas‘good’,‘mixed’or‘poor’.

Assessment of effectiveness

Tworeviewersindependentlycodedtheauthors’conclusionsregardingtheeffectoftheinterventionontimingofinitiationofantenatalcare,andindependentlyassessedandcodedtheevidenceofeffectiveness,takingintoaccountthestrengthsandlimitationsnoted in the GATE checklist.

Results

Overthreethousandcitationswerescreenedofwhichsixteenreports(eachrelatingtoadistinct intervention) met the inclusion criteria.

Fourteen(87%)ofthestudieswereconductedintheUS,1inAustraliaand1intheUK.

Thirteen of the sixteen included studies were observational cohort studies (10 were prospective,andthreewereretrospective;oneoftheretrospectivecohortstudiesalsoincludedapre-interventioncomparatorgroup);andthreewerebeforeandafterstudies.Allbutoneofthestudieswereassessedashaving‘poor’internalvalidity;onestudy(aretrospectivecohortstudy)wasratedashaving‘mixed’internalvalidity.

Twelvestudiesfocussedonspecificdisadvantagedorvulnerablesubgroupsofthepopulation.Thisincludedsixinterventionsthatweretargetedatand/orevaluatedinethnicminoritywomen,onethatfocussedonindigenousAustralianwomen,fourthattargetedteenagers,andonethatwasevaluatedinsubstanceabusingHIV-positivewomen.The remaining studies evaluated interventions in more generally socioeconomically disadvantagedpopulations.

Eleven studies evaluated interventions that involved outreach or other community-based services,andfivestudiesevaluatedinterventionsthatinvolvedalternativemodelsofclinic-basedantenatalcare.Themaincomponentsofeachinterventionandthetargetpopulationaresummarisedbelow.

Type of intervention Target population (number of studies)

Outreach or other community-based interventions

Layorparaprofessionalhomevisitingandsupport

Teenagers (n=2)

Socioeconomically disadvantaged women (n=1)

Linkworkers Ethnicminoritywomen/non-nativelanguagespeakers(n=1)

Mobile health clinics Socioeconomically disadvantaged women (n=1)

Multi-componentinterventions,includingtwoormoreofthefollowing:outreach,casemanagement,homevisiting,riskscreening,helpwithtransportationtoappointments,advocacyandsocialsupport

Ethnic minority women (n=5)

Indigenous women (n=1)


Type of intervention Target population (number of studies)

Interventions involving alternative models of clinic-based antenatal care

Teen clinics Teenagers (n=2)

Collaborative antenatal care Socioeconomically disadvantaged women (n=1)

Enhanced antenatal care Socioeconomically disadvantaged women (n=1)

Socioeconomicallydisadvantaged,HIV-positivesubstanceabusingwomen(n=1*)

*Interventiontargetedatsocioeconomicallydisadvantagedwomen;studyevaluatestheinterventioninHIV-positivesubstanceabusingwomen

Effectiveness

Outreach or other community-based interventions

Of the eleven studies evaluating the effect of outreach or other community-based interventionsonthetimingofinitiationofantenatalcare,onlyone(aparaprofessionalhome visiting intervention described below) was assessed as having adequate internal validity in relation to the estimated effect on the timing of initiation of antenatal care. The qualityofevidencerelatingtotheothercommunity-basedinterventionswaspoor.

Rogers and colleagues assessed the effectiveness of a home visiting intervention delivered byparaprofessionalwomen(‘resourcemothers’)onthetimingofinitiationofantenatalcareamongpregnantteenagers(agedlessthan18),usingaretrospectiveobservationaldesign.Theevaluationusedtwodifferentcomparisongroups,onedrawnfromdifferentbutbroadlysimilargeographicalareas,andtheseconddrawnfromadolescentswhoresidedintheinterventionareasbeforetheinterventionwasimplemented.Thestudywastheonlystudyincludedinthereviewwhichadjustedforpotentialconfoundingintheanalysisoftimingofinitiationofantenatalcare.Theevaluationreportedastatisticallysignificantincreaseintheproportionofinterventionteenagersinitiatingantenatalcarebeforethefourthmonthofpregnancyrelativetobothcomparatorgroups(interventiongroupvs.geographicalcomparatorgroup,45%vs.41%,adjustedoddsratio1.48(95%CI1.32,1.66);interventiongroupvs.‘pre-intervention’comparatorgroup,45%vs.40%,adjustedoddsratio1.39(95%CI1.16,1.66)).Theauthorsconcludedthatthestudydemonstratedabeneficialeffectonthetimingofinitiationofantenatalcare.Becauseofthepotentialforselectionbiasandnon-randomassignmentofparticipants,thereviewersconsideredthestudyinconclusivebutconsistentwithapossiblebeneficialeffect.

Interventions involving alternative models of clinic-based antenatal care

The quality of evidence relating to interventions involving alternative models of clinic-basedantenatalcarewaspoor.Allfiveoftheincludedstudieswereassessedashavingpoorinternalvalidityinrelationtotheestimatedinterventioneffect on the timing of initiation of antenatal care.

Conclusions

Inacomprehensivereviewofthepublishedliteratureontheeffectivenessofinterventionstoincreasetheearlyinitiationofantenatalcare,wefoundinsufficientevidenceofadequatequalitytomakeanyfirmrecommendations.However,oneincludedinterventionwasconsidered‘promising’;andthreeotherinterventionstrategieswereidentifiedthatwereconsideredpotentiallyrelevanttotheNHSandworthyoffurtherconsiderationandevaluation.


A systematic review of the effectiveness of interventions to increase the early initiation of antenatal care in

socially disadvantaged and vulnerable women

Thesystematicreviewdescribedinthisreportispartofaprogrammeofwork,commissionedbytheDepartmentofHealth,tostrengthentheevidencebaseoninterventionstoreduceinfantmortality,withaparticularfocusonreducinginequalitiesin infant mortality. The review focuses on interventions to increase the early initiation of comprehensiveantenatalcareinsociallydisadvantagedandvulnerablewomen.

Background1 Antenatalcareisconsideredtobeeffectiveinimprovingoutcomesforpregnantwomenand their infants.1 Evidence suggests that there is an association between under-utilisation ofantenatalcareandperinatalandinfantmortality.2 Early access to antenatal care is considered a key strategy in meeting targets to reduce inequalities in infant mortality in theUK,3andimprovingaccesstomaternityservicesisanongoingpriorityintheUK,4,5 witharecentGovernmentPSAtargetfocussingontheproportionofwomen‘booking’forantenatal care before 12 weeks.6,7

Initiationofantenatalcarewithinthefirsttrimesterisdesirable,withthemostrecentUKguidelines recommending initiation by 10 weeks gestation.1However,arecentUKsurveyfoundthatonly56%ofwomenhada‘booking’appointmentby12weeksgestation.8 Womenwhoinitiateantenatalcarelaterhaveareducedopportunitytofullybenefitfromtherangeofinterventionsofferedtopregnantwomen,forexampleearlyidentificationofriskfactorsforpre-eclampsiaandgestationaldiabetes,smokingcessationadvice,screeningforasymptomaticbacteriuria,1 and other screening tests offered to women in earlypregnancy.Thereisnoconsensusastowhatconstitutes‘late’booking.Asystematicreviewofsocialclass,ethnicityandantenatalcareattendance9 included studies in which thedefinitionoflateattendancevariedfrom14to20weeks;andareviewofbarrierstoaccess to antenatal care10founddefinitionsof‘latebooking’rangingfrom17to28weeks.

Onesystematicreviewhasconsideredtheassociationbetweensocio-demographicfactorsandattendanceforantenatalcareintheUK.9TheauthorsofthereviewidentifiedfiveUKstudieslookingatsocialclass,threeofwhichreportedanassociationbetweenmanualsocialclassandlateinitiationand/orunder-utilisationofantenatalcare.AllfourofthestudiesthattheyreviewedwhichconsideredethnicityreportedthatwomenofAsian origin were more likely to have delayed initiation of antenatal care. Other socio-demographicfactorsassociatedwithlateinitiationofantenatalcareintheUKincludeyoungerage,11smoking,11non-UKmaternalplaceofbirth,12 and single status (not married or cohabiting).12

Evidencefromotherdevelopedcountrieshasconfirmedassociationsbetweenlateinitiationofcareandlowersocio-economicstatus,13,14 belonging to an ethnic minoritygroup,14,15youngermaternalage,13–17smoking,15,17 and marital status.13,14,17 Inaddition,somestudieshavereportedassociationsbetweendelayedinitiationofcareandthefollowingsocio-demographicfactors:refugeestatus,18 low educational attainment,13,14,16,17,19highparity,13–15 alcohol use17andunplannedpregnancies.14,19

Both the characteristics of users and those of the health services themselves may affect access to care.20AmodeldevelopedbyCoopertoconceptualisebarrierstoequitablehealthcareforracialandethnicgroupsintheUSA(Figure1)21classifiespotentialbarriersintothreegroups:personal/familybarriers;structuralbarriers;andfinancialbarriers.Thismodelcanbeusedasastartingpointtoconceptualisebarrierstoantenatalcare,withsomebarriersrelatingtothe‘demandside’(forexamplehealthbeliefs,implicitorexplicitcostsofcare),andothersrelatingtothe‘supplyside’(e.g.qualityandavailabilityof services).22Significantly,thismodelwasdevisedforaUSsettingandfinancialbarriers


tocarearelikelytobedifferentintheUKsettingwithpubliclyfundedhealthcarewhichisfreeatthepointofaccessforthoseconsidered‘ordinarilyresident’.However,althoughthenumberofmigrantwomenintheUKineligibleforfreematernitycareisprobablysmall,thereisevidenceofconfusionamonghealthcarestaffevenwhenguidanceisclearthat women are entitled to National Health Service (NHS) care.23 For the majority of womenintheUK,financialcostsincurredinthereceiptofantenatalcarearelimitedto‘outofpocket’expensessuchastransportation,childcare,andpotentiallossofearnings,particularlyfor‘vulnerableworkers’suchashourlypaidcasualworkersandthoseinthe’informaleconomy’.However,thesemaybetangiblebarrierstocareforsomegroupsofwomen,e.g.thoselivinginruralareaswithoutadequatepublictransportation,thosecaringforotherchildrenandthoseininsecureemployment.24

Figure 1. Barriers to equitable healthcare for racial and ethnic groups (adapted from Cooper21)

Personal/family Structural Financial

Acceptability

Cultural

Language/literacy

Attitudes,beliefs

Preferences

Involvement in care

Health behaviour

Education/income

Health status

Availability

Appointments

How organized

Transportation

Eligibility

Insurance coverage

Reimbursement levels

Publicsupport(i.e.publicfunding)

Theprecisebarrierstocareexperiencedbywomenmayvaryaccordingtotheirsocio-demographiccharacteristics.NealeidentifiedarangeofbarriersexperiencedbyinjectingdrugusersintheUKastheyattempttoaccessgeneralhealthcareandsupportservices,andmanyofthesemaybesharedbyothervulnerableanddisadvantagedgroupsofwomenaddressedinthepresentreview.25 Although Neale’s study concluded that some barriersvariedaccordingtothesocio-demographiccharacteristicsofparticipants,others,forexamplestigmaandnegativeattitudesfromstaff,werereportedasexperiencedbyallinterviewees.Inaddition,asystematicreviewofaccesstoantenatalcareindevelopedcountrieshighlightsthe“varietyofsocio-demographic,economic,culturalandpersonalfactors” that affect the correlation between delayed or infrequent antenatal care and outcomes.10 A review of the qualitative literature by Lavender et al. suggests that for somehigh-riskmarginalisedwomen,simplyprovidingappropriateservicesislikelytobeinsufficientaswomenmaynotbehealthliterateandlackthepersonalautonomy,supportand/orabilitytomakeuseofthecarewhichismadeavailabletothem.26 These findingsaresupportedbyworklookingatwiderissuesofaccesstogeneralhealthcare,withonereviewemphasisingtheneedtoaddresssourcesofinequalitiesincare,with“keybarriers…unlikelytobeuniformacrosssectors,services,andgroupsofpeople”.22 Thisapproachissupportedbythenotionof‘candidacy’,asyntheticconstructdevelopedtodescribe“thewaysinwhichpeople’seligibilityformedicalattentionandinterventionisjointlynegotiatedbetweenindividualsandhealthservices”andusedtoemphasisethattheuse of health services requires considerable work by individuals.24,27 Dixon-Woods gives as anexampletheevidencesuggestingthatpeoplefrommoredeprivedbackgroundshavealowertake-upofpreventiveservices(the“inversepreventionlaw”28). Although this may beinpartattributabletostructuralbarriers,itmayalsoresultfromalackof“positiveconceptualisationofhealth”,andthetendencytomanagehealthanddiseaseasseriesofmajor and minor crises.27Thisexplanationmaybeparticularlyrelevanttothediscussionofantenatalcare,oftendescribedasoneoftheclassicexamplesofpreventivemedicine.29


Despitethepriorityplacedonearlyinitiationofantenatalcareandadevelopingbodyofevidenceonfactorsinfluencingcare,fewstudieshaveevaluatedstrategieslikelytoinfluencethetimingofinitiationofantenatalcare.Inourscopingwork,wedidnotidentifyanypublishedsystematicreviewsthatlookedexclusivelyatstrategiesforincreasingearlyinitiationofantenatalcare.However,wedididentifythreerelevantreviews,primarilyfocussingonperinataloutcomesbutwhichalsosynthesiseddataontheeffectsofincludedinterventions on the timing of antenatal care initiation:

One literature review evaluated changes in the delivery of antenatal care for •Australianindigenouswomen.Thereviewlookedatcareutilisationalongsidehealth/birthoutcomes.Tenevaluationswereincluded,fourofwhichreportedtimingofinitiation of antenatal care as an outcome measure.30

Oneliteraturereviewlookedattheeffectoflayhomevisitingonpregnancy•outcomes. The author synthesised the effect of included interventions on utilisation of antenatalcareforeightstudieswherethesedatawerereported.31

Oneliteraturereviewfocussedonevidenceaboutimprovingservicesfor•disadvantagedchildbearingwomenintheUK.Anumberofprimarystudiesandsystematicreviewswereidentified,reportingavarietyofdifferentoutcomesrelatingtotheperinatalperiod.Onlyoneintervention,focussingonethnicminoritywomen,reportedontimingofantenatalcarebooking.32

Aims of the review1.1

Thepurposeofthisreviewwastosystematicallyidentifyandevaluatetheevidencerelatingtotheeffectivenessofinterventions,relevantinthecontextoftheNHS,whichaimtoincreasetheearlyinitiationofcomprehensiveantenatalcareinsociallydisadvantaged and vulnerable women.

Definitions and scope of the review2 Weoperationalisedconceptsanddefinitionsasfollows.

Operational definition of comprehensive antenatal care2.1

Antenatalcarereferstopregnancy-relatedservicesprovidedbetweenconceptionandtheonsetoflabourencompassingmonitoringofthehealthstatusofthewomanandthefetus,provisionofmedicalandpsychosocialinterventionsandsupport,andhealthpromotion.33 Suchservicesaretypicallyprovidedasapackageofcare,whichweterm‘comprehensiveantenatalcare’,althoughsometimeselementsofantenatalcaremaybedeliveredseparately,forexamplehomevisitingprogrammestargetingpregnantwomen.34,35 In this reviewwefocusonthetimingofinitiationof‘comprehensiveantenatalcare’.

‘Early’ initiation of antenatal care2.2

CurrentUKguidelinesrecommendthatwomenreceivetheirbookingappointmentforantenatal care before 10 weeks.1Cut-offsusedtodefine‘late’bookingrangebetween14-28weeks,9,10withnoclearconsensusregardingtheoptimaldefinition.Twentyweeksmayberegardedasanuppercut-offpointbasedontheopportunitytoreceiveanultrasoundanomalyscanwithintherecommendedtimeperiod(18-20weeksintheUK1),althoughlatercut-offpoints(22weeks,26weeks,28weeks)arealsousedforthepurposesofmonitoringuptakeofcare.

Forthepurposesofthisreview,weconsideredtheeffectofinterventionsonthetimingofinitiationofantenatalcareuptoandincluding20weeksofgestation.


Types of intervention2.3

We were interested in any intervention which might be delivered to increase the early initiation of antenatal care by socially disadvantaged and vulnerable women. We envisaged that the majority of such interventions would be stand alone interventions or ‘outreach’ services attached to antenatal care services. However we also considered antenatal care serviceswithoutspecificoutreachserviceswithinthescopeofthereviewprovidedthatsome element of the intervention could be considered to address barriers to care.

NHS relevance2.4

InlinewiththeaimsoftheInfantMortalityProject,wedecidedtofocusoninterventionsthatwouldbeconsideredrelevantinthecontextoftheNHS.Inparticular,giventhepreponderanceofUS-basedresearchintheliterature,wespecificallywishedtoavoidtheinclusionofasubstantialvolumeofresearchrelatingtointerventionswhichprimarilyaddressedfinancialbarriersarisingfromlackofhealthcareinsuranceorinterventionsrelatingtostructuralorfinancialaspectsoftheUShealthcaresystemswhichwerenotapplicableintheUKcontext.Wewereunabletoidentifyanypublishedtypologyofinterventionsoraconceptualmodelwhichadequatelycapturedthisideaof‘NHSrelevance’.ThereforeweusedtheconceptsandcategoriesunderpinningthebarrierstohealthcareaccessmodeldevelopedbyCooper21 (discussed in Section 1 above) to operationaliseourinclusioncriteriarelatingtoNHSrelevance.

Usingthismodel,anyinterventionaddressingpersonal/familybarrierswasconsideredtobeofpotentialrelevancetotheNHS;interventionswhichwere‘primarily’structuralorfinancialwerenotconsideredrelevantunlesscomponentsoftheinterventionaddressedbarriersrelevanttowomenintheUKthatwerepotentiallytransferabletotheUKhealthcare setting.

Disadvantaged and vulnerable groups2.5

Wesoughtinterventionstargetingorevaluatedinthefollowinggroups.

Specificdisadvantagedandvulnerablegroupsofwomenatriskofaccessingantenatal•carelate,including:

Womeninprison ○Travellers ○Homeless women ○Asylum seekers and refugees ○Recently arrived migrants ○Otherimmigrantgroups ○Non-nativelanguagespeakers ○Victimsofabuse ○Womenwithmentalillness/mentalhealthproblems ○Women with learning disabilities ○Sex workers ○Victimsoffemalegenitalmutilation/cutting ○Teenagers ○WomenwhoareHIVpositive ○Substance users ○Alcohol misusers ○

Moregeneralgroupsofdisadvantagedwomen,including:•Women of low-socioeconomic status ○Womenlivingindeprivedareas ○Sociallydisadvantagedethnicminoritygroups ○


Methods3

Inclusion Criteria3.1

The following inclusion criteria were used:

Study design3.1.1

Norestrictionwasimposedonstudydesignotherthanthatthestudyhadtoincludeacontrolorcomparatorgroupandthestudymustbeanevaluationbroadlydesignedtocompareoutcomesintheinterventiongroupvs.thecontrol/comparatorgroup.Thusbothexperimentalandobservationalstudieswereeligibleforinclusion.

Population3.1.2

We required that the study evaluated the intervention in a socially disadvantaged or vulnerablepopulation,including,butnotlimitedtothegroupslistedinsection2.5above.Studiesthatevaluatedtheinterventioninamoregeneralpopulationbutprovidedsubgroupanalysisrelatingtorelevantsub-groupswerealsoeligibleforinclusion.

Intervention3.1.3

Wedidnotplaceanyrestrictiononthetypeofintervention.Werequiredonlythatstudiesreportedthetimingofinitiationofantenatalcareasanoutcomemeasure.

Comparator group3.1.4

We required that:

thestudyincludedacontrolorcomparatorgroupthatdidnotreceive,and/orhave•accessto,theintervention.

theinterventionandcomparatorgroupwereselectedusingthesameand/orsimilar•samplingframesi and that

theselectioncriteriaweresuchthatthetwogroupsweredrawnfrombroadlysimilar•populations.

Outcome measure3.1.5

Weincludedstudieswhichevaluatedtheeffectoftheinterventionontheproportionofwomeninitiatingcomprehensiveantenatalcarebyagivenweekormonthofpregnancyuptoandincluding20weeksofgestationorbeforethefifth,monthofpregnancy.

Somestudiesassessedtheeffectoftheinterventiononothercompositemeasuresofutilisationofantenatalcare,forexampletheKotelchuck Adequacy of Prenatal Care Utilization Index36 (a measure which takes account of both the timing of initiation of antenatalcareandthenumberofantenatalcarevisits,adjustedforthedurationofantenatal care). Such studies were eligible for inclusion if the timing of initiation of antenatalcarecomponentoftheindexwasreportedseparately.

Studiesthatreportedthetimingofinitiationofantenatalcareasabaselinecharacteristicwere excluded.

Language3.1.6

WeincludedonlyarticlespublishedinEnglish.

i Samplingframeswerenotconsideredcomparableif,forexample,oneincludedwomenwithnoantenatalcare(e.g.sampledfromabirthregister)andtheotherincludedonlywomenwithsomeantenatalcare(e.g.sampledfromaclinicpopulation)


Time period3.1.7

Modelsofantenatalcarehaveshiftedinrecentdecadesfrompredominantlyobstetrician-led/hospital-basedmodelsofcaretomorediversemodelswithgreaterinvolvementofmidwives,primarycarephysiciansandothersintheprovisionofantenatalcarefornon-highriskpregnancies.

Inordertofocusonmodelsofantenatalcarethatarerelevantinthecurrentcontext,weincludedonlystudiespublishedfrom1990onwards.

Geographical areas3.1.8

Welimitedthereviewtostudiescarriedoutinhighincomecountrieswithwelldevelopedhealthcare systems and relatively low infant mortality rates. We included interventions evaluatedinmembercountriesoftheOrganisationforEconomicCo-operationandDevelopment(OECD),exceptforMexicoandTurkey,bothofwhichhavemarkedlyhigherinfant mortality rates than the rest of the OECD.37

Types of publication3.1.9

Weincludedjournalarticlesreportingprimaryresearch,withorwithoutanabstract.

Exclusions3.2

InordertofocusoninterventionsrelevantinthecontextoftheNHS,weexcludedinterventions that related solely to:

theprovisionorextensionofhealthinsurancecoverage,orsimilar,forexample,•changesintheeligibilitycriteriaforMedicaid;

amodelofinsurancecoverageorreimbursement,forexample‘managedcare’or‘fee-•for-service’.

Wealsoexcludedinterventionsthatprimarilyaddressedotherbarrierstoantenatalcareaccessthatrelatedtostructuralorfinancialaspectsofthelocalhealthcaresystemwhichwerenotconsideredtoapplyinapredominantlygovernment-fundeduniversalhealthcaresystemsuchastheNHS(seediscussionofCooper’sbarriersmodel21 in Section 1 above).

Methods for identification of studies3.3

Overview of strategy to identify relevant studies3.3.1

Because of the diversity of the interventions which might be relevant and the absence of specificMESH/indextermsrelatingspecificallytouptakeofantenatalcare,weadopteda multi-stage strategy to identify relevant material. We initially carried out a range of scopingsearches,includinginternetsearches,toidentifypotentiallyeligibleinterventions.Basedonthis,wedevelopedalistofpotentiallyrelevanttextsearchtermsrelatingtospecificinterventionsandtypesofinterventions,whichwethenincorporated,togetherwithMESHandindexterms,inthesearchesrunonthemajorbibliographicdatabases(seesection3.3.2).Thetitlesandabstractsofstudiesidentifiedinthesesearcheswerescreened,asdescribedinsection3.4.1below.Duringscreeningonerevieweradditionallyflaggedstudiesrelatingtopotentiallyrelevantinterventions,irrespectiveofwhetherthestudymetthereviewinclusioncriteria.Basedontheflaggedinterventions,afurtherlistofinterventionswasdeveloped(listedinAnnexB)andthemajorbibliographicdatabaseswere again searched using these additional ‘free text’ terms relating to interventions of interest.


Bibliographic databases3.3.2

Thefollowingbibliographicdatabasesweresearchedinordertoidentifyreportsofprimarystudies using a combination of text terms and MESH headings relevant to the review (AnnexA).WesearchedforreportspublishedbetweenJanuary1990andApril2009,included in the following databases:

Medline (Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations and Ovid •MEDLINE(R)1950toPresent,searchedviatheOvidSPinterface)

Embase(EMBASE1988to2009Week15,searchedviatheOvidSPinterface)•

Cinahl (searched via the EBSCO interface)•

PsycINFO(PsycINFO1987toAprilWeek22009,searchedviatheOvidSPinterface)•

HMIC(HMICHealthManagementInformationConsortiumMarch2009,searchedvia•the OvidSP interface)

CENTRAL (searched via the Cochrane Library)•

Initial searches were carried out on 16thApril2009.

AfurtherroundofsearcheswascarriedoutusingthesedatabasesinMay2009,usingfreetextsearchstringsrelatingtoany‘named’interventionsidentifiedduringthefirstroundofscreening,asdescribedinsection3.3.1above.

Whereavailable,weappliedlimitsandfilterstorestrictthesearchresultsbypublicationyear(1990onwards),topic(humans),andlanguage(Englishlanguageonly).ThemainMedlinesearchwasadditionallyrestrictedonpublicationtypetoexcludeletters,news,editorials and commentaries.

AcopyofthemainMedlinesearchstrategyisprovidedinAnnexA.Alistofthe‘named’interventionsthatweincludedinthesecondroundofsearchesisgiveninAnnexB.Copiesof search strategies relating to other databases are available from the authors on request.

Other online searchable resources3.3.3

We searched the following databases through the Cochrane library interface to identify systematicreviews,guidelines,healthtechnologyassessmentsandeconomicevaluationsdealingwithaccesstoantenatalcareorrelatedtopics:

Cochrane Database of Systematic Reviews•

Database of Abstracts of Reviews of Effects (DARE)•

Health Technology Assessment Database•

These databases were searched on 28thApril2009.ThestrategyusedtosearchthesedatabaseswasidenticaltothatusedtosearchCENTRAL,andusedacombinationoftextterms and MESH headings relevant to the review.

Weadditionallysearchedthefollowingspecialistdatabasesandonlineresourcesinordertoidentifyanyfurtherprimaryreports,orguidelines,reviewsandreportswithrelevantcitations:

The National Guideline Clearing House•

The National Library for Health•

The National Institute for Health Research Service Delivery and Organisation •Programme

OpenSIGLE•

TRoPHI•

TheHealthDevelopmentAgency(HDA)•

The National Institute for Health and Clinical Excellence (NICE)•


These databases and online resources were searched between 23rd and 30thApril2009.Whereasearchfacilityexistedwithinaparticulardatabase,abasicsearchwasconductedusing text terms relevant to the review. The reference lists of the relevant systematic reviews,guidelines,etc.werecheckedtoidentifyanyadditionaleligiblestudies.

Items identified in scoping exercise and antenatal care review3.3.4

Weincludedrelevantstudiesidentifiedduringinitialscopingworkdescribedabove(section3.3.1)Additionally,asmallnumberofitemsevaluatingrelevantinterventionswereidentifiedduringtheconductofarelatedsystematicreviewfocussingonantenatalcareinterventions.38Theitemswereincludedforscreeningalongsidematerialidentifiedfromothersourcestoensurethatinclusioncriteriawereappliedconsistently.

Reference lists and citations3.3.5

Followingthefulltextscreeningstage,thereferencelistsofallincludedstudieswerecheckedandfulltextversionsofanypossiblyrelevantcitationswereretrievedandscreened. We also searched the Science Citation Index via the Web of Science to recover anyrelevantpapersthatcitedanyitemsalreadyscreenedaseligibleforinclusion.

Review methods3.4

Screening3.4.1

Forthepurposesofscreening,theeligibilitycriteriadescribedinsection3.1and3.2abovewere reformulated as a set of exclusion criteria as shown in Table 1.

Abstract screening3.4.1.1

Titlesandabstracts(whereavailable)werescreenedindependentlybytworeviewersusing the exclusion criteria listed in Table 1. Articles were included for full-text review if eitherofthereviewersconsideredthestudypotentiallyeligibleonthebasisofthetitle/abstract.


Table 1. Exclusion criteria

Stage 1: Abstract/title screening Stage 2: Full-text screening

Stage 1 criteria PLUS:

General Not English language•

Notprimaryresearch•

Noteligiblepublicationtype•Not journal article – e.g. dissertation, book, conference abstract

Population Not conducted in an eligible •OECD country

Notpregnantwomenorstudy•populationnotrelevant

Intervention No relevant intervention•Study does not evaluate any form of intervention OR evaluated intervention could not reasonably be expected to influence the timing of antenatal care initiation

Ineligible intervention•Study intervention relates only to the provision or extension of health insurance coverage or similar, OR the study intervention relates only to the model of insurance coverage/ reimbursement, OR study relates only to other non-relevant structural or financial interventions, for example healthcare fees, cost of malpractice suits, liability cover etc.

Comparator Nocomparator/controlgroup• Controlgroupnoteligible•Inappropriate comparator/control group i.e. the comparator/control group is not drawn from a population of interest and/or the intervention and control group are drawn from different and non-comparable populations

Outcome No relevant outcome• No relevant outcome•Timing of initiation of antenatal care reported but not an outcome measure


Stage 1: Abstract/title screening Stage 2: Full-text screening

Stage 1 criteria PLUS:

Other Not an effectiveness evaluation•Study does not report an effectiveness evaluation of a relevant intervention with an eligible control group and a relevant outcome measure reported

Selection criteria for study •groupsnotappropriateFor example, - RCTs where women were randomised after entry into antenatal care - Studies where inclusion/exclusion criteria were based on the timing of initiation of antenatal care, e.g. studies which excluded ‘late bookers’ - Studies where the intervention and control/comparator groups were sampled from non-comparable sampling frames (e.g. antenatal care clinic records vs. birth records)

Full text screening3.4.1.2

Thefulltextarticlesofallitemsincludedattheabstract/titlescreeningstagewereretrievedandscreenedindependentlybytworeviewersusingtheexclusioncriteriausedpreviouslyandadditionalmorespecificcriteria(Table1).Reviewerswereaskedtousetheirjudgementincaseswhereanitemwasnotexplicitlyreported.Althoughwedidnotrestrictbyinterventiontype,studieswereexcludedatthetitle/abstractstageiftherewasnoevidencethattimingofinitiationofantenatalcarewasreportedandreviewersconsideredthattheinterventioncouldnotfeasiblybeexpectedtoinfluencetimingofinitiationofantenatalcare.Whereitwasnotexplicitlystatedthatthetimingofinitiationofantenatalcarewasastudyoutcome,reviewerswereaskedtoassesswhetherthiswasreportedasabaselinecharacteristicorasanoutcomemeasure.

Wheretherewaslackofagreementbetweenthereviewerstheopinionofathirdreviewerwas sought and a decision reached following discussion. It was found that the reviewers were sometimes unable to reach a clear consensus as to whether the timing of initiation ofantenatalcarewasanoutcomemeasure;toavoidtheexclusionofpotentiallyrelevantmaterial,thesestudieswereincludedbutthisaspectofthestudywascodedas‘unclear’.

Quality assessment3.4.2

An assessment of internal validity was carried out using the GATE checklist.39 Two reviewersindependentlyassessedeachstudy,andawardedanoverallgrade:++ Good:wellreportedandreliable;

+ Mixed:someweaknessesbutinsufficienttohaveanimportanteffectonusefulnessofstudy;

- Poor:studynotreliable,notuseful.


Wherethetwoassessmentsdidnotagree,theopinionofathirdreviewerwassoughtandafinalgradewasassignedfollowingdiscussion.

Foranalysispurposes,studiesassessedas‘mixed’or‘good’werecombinedtoprovidean‘adequate’ category.

PriortoundertakingthestudyGATEassessments,reviewerscompletedanddiscussedaminimumoffive‘trainingassessments’toensurethatthetoolwasbeingcorrectlyandconsistentlyapplied.

Data extraction3.4.3

AdataextractionandcodingformwasdevelopedandloadedintoEppi-Reviewer,40 customisedsoftwaredesignedtomanagescreening,dataextractionandanalysisforsystematic reviews.

Basicdescriptivedatawerecodedbyonerevieweronly;otherinformation,e.g.relatingtotheaims,studydesign,resultsandconclusions(assessmentofeffectiveness,seebelow)oftheevaluationwasindependentlycodedbytworeviewersandresultscompared.Discrepancieswereresolvedbydiscussionwithathirdreviewerconsultedifnecessary.

Assessment of effectiveness3.4.4

Authors’ conclusions3.4.4.1

Authors’ conclusions on the effect of the intervention on the timing of initiation of antenatalcarewereindependentlyassessedbytworeviewersandcodedasfollows:

+ Statisticallysignificantbeneficialeffect

(+) Effectconsistentwithbeneficialeffectbuteffectnotstatisticallysignificantand/orcautiousinterpretationoffindingsuggested

X Noevidenceofbeneficialeffect

0 No conclusion stated

Wherethereviewersdisagreed,athirdreviewerassessedthestudyandadecisionwasreached following discussion.

Reviewers’ assessment of effectiveness3.4.4.2

Tworeviewersassessedandindependentlycodedtheevidenceofeffectiveness,takingintoaccountthestrengthsandlimitationsnotedintheGATEchecklist,withinputfromathirdreviewerasdescribedpreviously.Studiesweregradedusingthefollowingcategories:

+ Studydemonstratesabeneficialeffect

(+?) Studyinconclusivebutmaydemonstrateabeneficialeffect

X Studydoesnotprovideconvincingevidenceofabeneficialeffect

Studiesratedashavingpoorinternalvalidity(i.e.GATEqualityassessment‘Poor:studynotreliable,notuseful’)werenotconsideredfurther.

Results4 ThenumberofitemsincludedateachstageofthereviewispresentedinFigure2.


Figure 2. Screening process

Citations identified from initial search of major

bibliographic databasesn=3069

Stage 2 full-text screening

n=125

Stage 1 abstract/title screeningn=2100

Duplicatesn=1166

Excluded on abstract/title

n=1975

Excluded on full-text review

n=109

Included in reviewn=16

Citations identified from other sources

n=197


Initialsearchesofthemajorbibliographicdatabasesidentified3069citations,ofwhich1062wereduplicates.Afurther197citationswereidentifiedfromothersources:fromnamedinterventionsearches,scopingsearches,theantenatalcaresystematicreview,andfromcheckingthereferencelistsandcitationsofstudiesidentifiedforinclusion.Onehundredandfourofthiscitationsidentifiedfromothersourceswereexcludedasduplicates.Overall,2100itemswerescreenedontitle/abstract(stage1),ofwhich1975wereexcluded.Ofthe125progressingtofulltextreview,109wereexcludedasaresultof full-text screening (stage 2). Further information about reasons for exclusion are presentedinTable2.

Table 2. Reasons for exclusion

Reason for exclusion

Excluded at stage 1: Abstract/title screening (n=1975)

Excluded at stage 2: Full-text screening (n=109)

General Notprimaryresearch

Noteligiblepublicationtype

259 52

16 3

Population Not conducted in an eligible OECD country

Notpregnantwomenorstudypopulationnotrelevant

627

149

0

0

Intervention No relevant intervention

Ineligible intervention

827 35

0 1

Comparator Nocomparator/controlgroup

Controlgroupnoteligible

5 0

20 11

Outcome No relevant outcome 18 49

Other Not an effectiveness evaluation

Selectioncriteriaforstudygroupsnotappropriate

3

0

4

5

Overview of included studies4.1

Weidentified16eligibleevaluationsrelatingto16distinctinterventions.Thefollowingsectionsdescribethese16primarystudies.

Countries4.1.1

FourteenoftheincludedstudieswereconductedintheUSA,onewascarriedoutinAustralia,41andoneintheUK.42

Year of publication/study4.1.2

Thesearchesidentifiedstudiespublishedbetween1990and2009.Themostrecentstudyincludedinthereviewwaspublishedin2007.Themajorityofstudieswerepublishedbetween1996and2001.ThedistributionofstudiesbyyearofpublicationispresentedinFigure 3.


Figure 3. Year of publication of included studies

0

1

2

3

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

Num

ber o

f stu

dies

Year of publication

Fourteenstudiesexplicitlystatedthestudytimeperiodfortheevaluation.Ofthesefourteenstudies,fourwerecompletedbefore1990,fivewerecompletedbefore1995,onewascompletedbefore2000,andthreewerecompletedbefore2005.Fortheremainingthreestudies,thetimeperiodoftheevaluationwasnotstatedorwasunclear.

Study design4.1.3

All the included evaluations used observational study designs. Two were before and after studieswithoutacontemporaneouscomparisongroup,43,44 and one was a before and after studywhichincludedacontemporaneouscomparatorgroupbutdidnotreportdataonthetimingofinitiationofantenatalcareforthisgroup.41Onestudywasaretrospectiveobservationalcohortstudywithanadditionalpre-interventioncomparatorgroup.45 The remaining12studieswerecohortstudies,ofwhichninewereclassifiedasretrospectiveandthreeasprospective.Oneoftheseincludedamatchedcomparatorgroup.46

Outcome measure4.1.4

Justoverhalfofthestudies(n=9)reportedtheproportionofwomeninitiatingcareinthefirsttrimester.41,46–53Twostudiesreportedmeasuresbasedoninitiationofantenatalcareby 12 weeks42orby14weeks,54andfivestudiesreportedthemonthofpregnancywhenantenatal care started (before the fourth44,45,55andfifthmonth43,56ofpregnancy).Thesource of data on gestation at initiation of antenatal care varied: six studies (all US-based) usedinformationrecordedonthebirthcertificate,43,45–47,50,56 three used clinical records only,41,42,44andone,whichrecruitedrecipientsofthe“SpecialSupplementalFoodProgramforWomen,InfantsandChildren”(WIC)services,usedtheWICrecordsasthesourceofdata.48Intheremainingsixstudies,thesourceofinformationongestationatinitiationof antenatal care was not clearly stated.49,51–55Onlytwostudiesexplicitlyreportedtheprocessbywhichgestationalagewasascertained.41,54


Quality4.1.5

Inter-raterreliabilityoftheGATEtoolwaslow(Kappa=0.18),with25%ofinitialassessmentsdiscordant(n=4).Fifteenoutofsixteenstudiesweregivenafinalratingof“poor”,andonestudywasratedashaving“mixed”internalvalidity.45Thepoorinternalvalidityoftheincludedstudiespartlyreflectedtheinclusionofanumberofstudiesinwhichinitiationofantenatalcarewasnottheprimaryfocusoftheevaluation.Thereviewers assessed internal validity in relation to the evaluation of effects on the timing ofinitiationofantenatalcare;incaseswherethiswasnottheprimarystudyoutcome,thisratingdoesnotnecessarilyreflectthevalidityofestimatedeffectsonotherstudyoutcomes.

Themostcommonlyreportedflaw(15studies)wasalackofadjustmentforpotentialconfounding in the analysis of the effect of the intervention on the timing of initiation of antenatalcare(somestudiesreportedadjustedanalysesforotheroutcomemeasures).Thiswasaseriousproblemasmanystudiesalsoreportedsignificantbaselinedifferencesbetweentheinterventionandcomparatorgroups,oftenaresultoftheinterventiontargetinggroupswithahigherriskprofile.SomekeyflawsidentifiedintheincludedstudiesarereportedinTable3.

Table 3. Major flaws identified in included studies

FlawNumber of studies affected*

Reporting of the study

Nodatapresentedonbaselinecharacteristicsbyintervention/comparatorstatus

3

Insufficientdatapresentedonbaselinecharacteristicsbyintervention/comparatorstatus

1

Outcomedatapresentedonlyingraphicalform,nonumericresultsprovided

2

Design of the study

Interventionandcontrolgroupsknown to differ at baseline with regard to importantcharacteristics,andnoadjustmentforknowndifferencesattheanalysis stage

8

Interventionandcontrolgroups likely to differ at baseline with regard to importantcharacteristics(insufficientdatapresentedtoassess),andnoadjustment for likely differences at the analysis stage

2

Smallsamplesize(n=<200) 4

Noprotectionagainstsecularchanges(beforeandafterstudywithoutcontemporaneouscomparisongroup)

3

Atleastonecomparatorgroupincludeswomenwhomayhavereceivedthe intervention under study (contamination)

2

Analysis of the study

Noadjustmentforpotentialconfoundinginanalysisoftimingofinitiationof antenatal care

15

Inappropriateanalysismethod(unmatchedanalysisformatcheddesign) 1

*Numbersdonotaddupton=15,moststudieshadmultipleflaws


Assessment of whether timing of initiation of antenatal care was 4.1.6 an outcome measure

Overall,eightstudieswereconsideredtoclearlyreporttimingofinitiationofantenatalcare as an outcome measure. In the remaining eight studies it was considered unclear as towhethertimingofinitiationofantenatalcarewasreportedasanoutcomemeasureasopposedtoabaselinecharacteristic.

Interventions studied4.2

Intervention recipients/target populations4.2.1

Bydefinition,allincludedstudiescoveredinterventionsthatweretargetedatand/orevaluatedinoneormoreofthedisadvantagedorvulnerablegroupslistedpreviously.Twelvestudiesfocussedonspecificsubgroupsofinterest.Thisincludedsixinterventionsthatweretargetedatand/orevaluatedinethnicminoritywomen(“womenfrom‘minority’backgrounds”,46Mexican-American,43African-American,47,50,53 Asian-British42),onethatfocussedonindigenousAustralianwomen,41fourthattargetedteenagers,44,45,51,55 andonethatwasevaluatedinsubstanceabusingHIV-positivewomen.56 Four studies coveredinterventionstargetedatand/orevaluatedinmoregenerallysocioeconomicallydisadvantagedpopulations,49,52,54oneofwhichsimplydescribedtheinterventiontobetargeted at “at risk families”.48

Intervention content4.2.2

The16includedinterventionswerebroadlyclassifiedaccordingtowhethertheywereoutreach or other community-based interventions (11 studies41–43,45–50,53,55);orwhetherthey were interventions involving alternative models of clinic-based antenatal care (5 studies44,51,52,54,56).

Outreach or other community-based interventions4.2.2.1

Eleven studies evaluated outreach or other community-based interventions. Three of theseinterventionsconsistedprimarilyofsocialsupportand/orhomevisitsdeliveredbyparaprofessionalorlaywomen.45,48,55Ofthesestudies,twoevaluatedinterventionsbasedontheconceptof‘resourcemothers’–trainedparaprofessionalwomenrecruitedfromthelocalcommunity-providingsupporttopregnantteenagers.45,55 The third intervention encompassedhomevisitingforsocioeconomicallydisadvantaged“atrisk”families.48 One interventionconsistedoftheprovisionof‘linkworkers’inprimarycareandantenatalcaresettings,42 and in another study the intervention was a mobile health clinic offering basic antenatal services.49Theremainingsixstudiesallevaluatedmulti-componentinterventionsincludingtwoormoreofthefollowingcomponents:outreach,casemanagement,homevisiting,riskscreening,helpwithtransportationtoappointments,advocacyandsocialsupport.41,43,46,47,50,53 Five of the interventions 41,43,46,47,50 involved lay workersorparaprofessionalstaffindigenoustothetargetedcommunity.

Lay or paraprofessional home visiting and support

RogersandcolleaguesevaluatedtheimpactofaResource Mothers Program (RMP) in a sampleofruralandmoderatelyurbancountiesinSouthCarolina,USA.45 The intervention wasdeliveredbyresourcemothers(paraprofessionalwomenwhoprovidedsocialsupportthrough home visits). These women were recruited from the local community and received threeweeksofintensivetrainingonarangeofsubjectsincludingpregnancyandinfantcare,nutritionandcommunicationskills.Pregnantteenageparticipants(<18years),whowerepredominantlyBlack,wererecruitedthroughoutreachactivitiesorthroughpeer-referral or referral from other agencies such as the Special Supplemental Food Program for Women, Infants and Children (WIC),schools,antenatalcareclinicsandchurches.Theresourcemothersprovided“supportive,educationalhomevisits”andhelpedtheteenager“usethehealthcaresystem”.Afterenrolment,teenagerswerevisitedmonthlyduringpregnancy,afterdeliveryinhospital,andmonthlyforthefirstyearoftheirinfant’slife.


“Each visit was structured, with specific goals and learning objectives. Prenatally, emphasis was on the need for early and regular prenatal care and reduction of risk factors, such as smoking, drug use, and poor nutrition.”

“Resource mothers facilitated the teenagers’ use of prenatal care and support services by following up on any missed appointments, arranging transportation, and assisting with referrals to community and health services. The resource mother acted as an advocate for the participant by bringing attention to her needs within health and community agencies.” 45

JulnesandcolleaguesevaluatedanotherResource Mothers Program,basedinNorfolk,Virginia,USA.55Theinterventionwastargetedatpregnantteens(<18years)withcertainriskfactors:youngmaternalage,black,residingintargetneighbourhoodswithlowfamilyincomelevels,lessthanahighschooleducation,andnopriorpregnancies.AswiththeinterventionevaluatedbyRogers,resourcemotherswererecruitedfromthecommunityandprovidedwithintensivetrainingtoenablethemtosupportpregnantteenagersfromdisadvantaged backgrounds.

“This program utilizes “resource mothers” to reach out to adolescents considered at high risk for inadequate prenatal care and poor pregnancy outcomes. A resource mother is a lay person – often indigenous to the culture of the adolescents – trained to assist adolescent parents and their families with the non-medical dimensions of pregnancy and child care. The resource mother is responsible for recruiting teens for the program, encouraging them to get prenatal care, providing practical assistance to the teens and their families, and acting as a liaison between the teens and the relevant public agencies.” 55

Daaleman evaluated the Kansas Healthy Start Home Visiting (HSHV) Programme.48 This programmewasdesignedtoenableat-riskfamiliestobecomehealthierandmoreself-sufficientbyimprovingaccesstoearlyinterventionservices.Thisevaluationwasdesignedtoinvestigatewhetherpriorexposuretothisprogramme(i.e.beforepregnancy)hadaneffectontheuseofantenatalcareinthecurrentpregnancy.TheevaluationwasconductedusingasmallsampleofmultiparouswomeninreceiptofWICservices.HSHVwasacommunity-basedlayhomevisitingprogramme,availableto“allpregnantwomen,infants,adoptivefamilies,andfamilieswhohavelostanewborn”.Participantswerereferredbytheirphysician,careproviderorsocialserviceagency.

“The home visitor is an experienced parent with a minimum of a high school diploma or GED, who has undergone an orientation to home visiting under the supervision of a public health nurse. The role of the home visitor is to provide education, support, resource information and referrals to the family, in addition to screening for any current or potential problems. No childcare or transportation services are provided by the home visitor. All visits are reviewed with a public health nurse to assess for any necessary follow-up or referral.” 48

Linkworkers

Mason evaluated the Asian Mother and BabyprojectinLeicester,UK.42Theproject(whichwaspartiallyhospital-based)involvedeightAsianlinkworkersbasedacrossthetwomaincity maternity units (two linkworkers on each site) and four selected GP surgeries (one linkworkerateachsurgery).GPpracticeswereselectedfromthosewhichhadatleastonegeneralpractitionernotonthe‘obstetriclist’i. The linkworkers were “women aged between 20and45whowereabletospeakfluentEnglishandatleastoneAsianlanguage”.Thelinkworkers“workedalongsidehealthprofessionals,inbothhospitalandcommunityantenatalclinics,as‘facilitators’and‘interpreters’whilealsofulfillinganeducativerole”.Theaimoftheinterventionwastoimprovebirthoutcomes,aidcommunicationwithprofessionals,andtoimparthealtheducation.

i Aregisterofgeneralpractitionerswhohavecompletedaspecifiedleveloftraininginobstetricsandgynaecology


Mobile health clinics

OnestudyevaluatedamobilehealthclinicforwomeninCalifornia,USA.49 The Women’s Health Vanwasstaffedbyanobstetrician-gynaecologistandnursepractitionerandprovidedavarietyofservicesforwomen,includingpregnancytesting,sexuallytransmittedinfection(STI)screening,breastexamsandcontraceptiveservices.Thestaffonthevanwerebilingual(English/Spanish)andthehealtheducationliteraturetheyprovidedwasavailableinbothlanguages.Thevanaimed“toaddressbarrierstohealthcareaccesssuchaslanguage,transportation,andcostforundocumentedimmigrantsandthe uninsured”. Two days a week the van travelled to low-income neighbourhoods and providedfreewalk-inorappointmentservicestolocalwomen.

“Women with positive urine pregnancy tests received a dating ultrasound on the van, initial prenatal care, counseling regarding healthy pregnancy, and are given a packet of information and prenatal vitamins. The van acts as a bridging device as the women are then referred to local community clinics for further prenatal visits” 49

Multi-component interventions

Cramerandcolleaguesreportanevaluationofacommunity-basedantenatalcareprogrammecalledOmaha Healthy Start,implementedinNebraska,USA.47 The setting wasspecificcensustractsinDouglasCounty,where46%ofthepopulationwereBlack.Theinterventionwasdesignedto“reducelocalracialdisparitiesinbirthoutcomes”.The intervention was delivered by outreach workers (indigenous to the targeted Black community),socialworkersandpublichealthnurses.Outreachworkerswereresponsibleforrecruitingpregnantwomentotheintervention,achievedthroughcommunityoutreachamong“localchurches,clinics,socialserviceagencies,communitygroups,communityleaders,andbusinesses”.Onceenrolledintheprogramme,womenwereassignedacasemanager(asocialworkerorpublichealthnurse)whoprovided“weeklycontact,throughhomevisits,officevisits,ortelephonecalls”.Casemanagersscheduledmedicalandothervisits,helpedtoarrangetransportationtoappointments,andscreenedandreferredparticipantsforriskfactors.CasemanagersalsodeliveredantenataleducationaccordingtotheprogrammedevelopedbytheNationalHealthy Startprogramme.

An effectiveness evaluation of the Rural Oregon Minority Prenatal Program (ROMPP) is reportedbyThompsonetal.43Thisinterventionwastargetedatlow-income,Mexican-AmericanwomenatriskofpoorbirthoutcomesinaruralOregoncommunityintheUSA.ROMPPattemptedtodeliver“culturallyappropriatecare,outreach,nursingcasemanagement,andhomevisitation”tothisgroupofwomen,manywhomwereundocumented immigrants and ineligible for Medicaid. The intervention was delivered byacommunityhealthnurse/casemanagerandoutreachworkers.Thecommunityhealthnurse/casemanager“wasresponsibleforassessment,planning,coordinationandevaluationofnursingcare”.Aswellasfacilitatingaccesstoantenatalcare,thenurse/casemanager was able to refer and liaise with other community services (e.g. WIC) as needed. TheoutreachworkerwasdrawnfromthelocalMexican-Americanfarmworkercommunity,and“wasresponsibleforcase-findingandrecruitment,follow-uptoensurecontinuityofcareandreducesocialisolation,andadvocacytolowerbarriersandincreasetheacceptabilityandaccessibilityofcare”.ROMPPreferredwomentothirdpartysourcesoffinancialhelpwithcarecosts,andnegotiatedpaymentarrangementsforwomenfundingtheirowncare.MostROMPPvisitsoccurredinparticipants’homes,withthenumberofoverallvisitsdependentontheneedsofthewomen.Theoutreachworkerprovidedtransportationtoantenatalcareappointmentsandinterpretingserviceswherenecessary.

WillisandcolleaguesreportanevaluationoftheBlack Infant Health (BIH)programme,targetedatAfrican-AmericawomenlivinginCalifornia,USA.53 BIH included “augmented servicesduringtheprenatalperiod,servicesdesignedspecificallyforAfrican-Americanwomen,outreachandtracking,office-basedservicesenhancedbytelephoneandin-homecontacts,andpreserviceriskscreening”.Itwasseparatetoantenatalcare“butconsistentlyenabledandsupportedclientswithprenatalcareentryandcontinuance”.ExactservicesprovidedbyBIHvariedbyprogrammesite.Allprogrammesitesimplementedthe‘PrenatalCareOutreach’model:


“The Prenatal Care Outreach model utilizes community health outreach workers to conduct intensive outreach to identify and link pregnant African-American women to BIH, general prenatal care, and other appropriate services” 53

Uptothreeothermodelswereimplementedaspartoftheprogramme,dependingon the results of local needs assessment. These models of care included the “Case Management”model(“utilizespublichealthnursestoconducthomevisitsforthepurposeofassessments,referrals,provisionandcoordinationofservices,monitoring,andfollow-up”),the“SocialSupportandEmpowerment”model,andthe“RoleofMen”model.

An evaluation of the Minority Health Coalitions Early Pregnancy Project was carried out byJewellandRussell.46Theintervention,implementedinIndiana,USA,evolvedfromtheIndianaMinorityHealthCoalitionswiderbriefto“eliminatehealthdisparitiesforracialandethnicminorities”.Theinterventionthatformsthefocusoftheirreportwasdesignedtoincreaseaccesstoearlyantenatalcare.Theprojectaimedto“eliminateculturalbarrierstocare”.

“The cultural aspect of care was emphasized in the projects as demonstrated by the use of minority professional and paraprofessional staff and the monitoring of the projects by the minority health coalitions boards [...] Staff provided social support in varying ways from individual support via contact with mothers in the project offices and on home visits, to group support by facilitating linkages of social support with significant others and holding support group meetings of the project mothers. Other interventions included referrals to community services, health education and transportation. The staff also provided advocacy for the mothers if barriers occurred in navigating health and social service systems in their communities.” 46

The Maternal Infant Health Advocate ServiceprogrammewasimplementedintheurbanareaofFlint,GenesseeCountyinMichigan,USA.50 Hunte and colleagues conducted an effectivenessevaluationofthisintervention,targetedat,andevaluatedamong,African-Americanwomen.Theauthorsreporttheobjectivesoftheinterventionasfollows:

“1) to identify pregnant African-American women early in their pregnancies; 2) to assist identified participants in navigating the prenatal care system; 3) to identify resources that assure services are adequate to reduce the stress associated with health barriers; and 4) to engage participants in other activities that assist in addressing issues of race and ethnicity as they relate to infant mortality.” 50

Participantsintheintervention(clients)wereidentifiedthroughself-referral,advocatecase-finding,andthroughreferralfromotherservicesandsettings(clinics,WIC,localhealthdepartmentsetc.).

“Upon entering the MIHAS program, clients meet face-to-face with their advocates to set specific goals to be addressed during their enrolment. While enrolled in the program all clients must be actively working towards their goals.” 50

Becauseclientsreportedthatphysicians“talked-down”tothem,advocatesalsoaccompaniedwomentoantenatalandpostnatalvisits,andinfantcheck-ups.

“[Advocates also provide] supportive services ranging from providing assistance when seeking employment, and help with school enrolment, to continuing their educational goals. Poor reading skills among many of the clients is a known barrier therefore advocates often accompany their clients to provide assistance and support with filling out necessary paperwork.” 50

TheinterventionevaluatedbyMackerrasandcolleagues“hadspecificgoalstoincreaseinfantbirthweightsbyearlierattendanceforantenatalcareandimprovedmaternalweight status”.41Theintervention,namedStrong Women Strong Babies Strong Culture wasevaluatedinruralAboriginalcommunitiesintheNorthernTerritory,Australia.TheinterventionwasdevelopedinconsultationwiththelocalAboriginalpopulation,andlaywomen indigenous to the community were trained as “Strong Women Workers” (SWWs).


“…a well respected Aboriginal woman was employed to develop the project. She worked with women selected by the communities (the SWWs) to implement a program that included traditional cultural practices related to childbirth as well as informing pregnant women about Western health and medical practices related to pregnancy and encouraging greater use of antenatal health care.” 41

Theinterventionalsotargetedwomennotyetpregnant,andthosewomenwhowerepregnantbutnotyetreceivingantenatalcare.

Alternative models of clinic-based antenatal care4.2.2.2

Fivestudiesreportedinterventionsthatinvolvedalternativemodelsofclinic-basedantenatalcare.Twoofthereportedinterventionswereteenantenatalclinics,44,51 one studyevaluatedacollaborativecareinitiative,54andtworeportedevaluationsofenhancedantenatal care services.56,52

Teen pregnancy clinics

MartinandcolleaguesevaluatedtheimplementationofateenpregnancyclinicinCincinnati,Ohio,USA.44 The evaluation was conducted through a small before and afterstudy.Theclinicwassetuptoprovidecomprehensiveantenatalcaretopregnantteenagerswhowerepreviouslyonlyabletoreceivenon-specificcarethroughthetraditional antenatal clinic.

“The operational objectives of the teen pregnancy clinic were to increase compliance among teen patients receiving care through GHA in attending prenatal appointments, educational classes and postpartum checkups. Some broader objectives of the clinic included reducing the number of teens who deliver low birth weight and premature infants, improving neonatal outcomes, decreasing the number of repeat pregnancies, decreasing the incidence of sexually transmitted diseases, and ensuring compliance with contraceptive care.” 44

Allteenagersparticipatingintheevaluationwereaged<18yearsandthemajorityhadtheircarefundedviaprivatehealthinsurance.

AnotherevaluationofteenpregnancyclinicswasconductedbyMorrisandcolleaguesinTexas,USA.51Thesettingwasapublichealthclinicservingamulti-ethniclow-incomepopulation,themajorityofwhomweremedicallyindigent.Theclinicwasdesignedforpregnantteenagers<18yearsandprovided:

“…general monitoring of the course of pregnancy, in addition to special emphasis on educational, social and nutritional support. The care was provided by a team of nurses, physician assistants, obstetrician-gynecologist residents, a social worker, and a nutritionist.” 51

Collaborative antenatal care

MvulaandMillerevaluatedacollaborativeantenatalcareprogrammeinLouisiana,USA.54 Theclinic,Neighbourhood Pregnancy Care, was situated next to low-income housing projectsinNewOrleansandprovidedcontraceptiveservicesalongsideantenatalcare.Theclinicfocussedon“continuityofprenatalcarebyspecificproviders,individualizedperinataleducation,andnursingcasemanagement...”.Servicesweredeliveredbyteamsofobstetriciansand‘advancedpracticenurses’(clinicalnursespecialists,nursepractitioners,andnursemidwives).Tomaximisecompliance“patientsareremindedthedaybeforescheduledappointments”.

Enhanced antenatal care services

One study evaluated the Prenatal Care Assistance Program (PCAP).56 PCAP was a combined state-federal intervention delivered through selected Medicaid clinics in NewYorkState,USA.ClinicswereeligibletobepartofthePCAPiftheydeliveredspecificservicesalongside“comprehensiveprenatal,diagnosisandtreatmentservices”.


Thespecificservicesincludedcarecoordination,referralstootherservices(HIVmanagement,mentalhealthservices,andsubstanceabuseprogrammes),andhealthandnutritioneducation.ClinicsdesignatedaspartofPCAPwereabletobillforantenatalandpostpartumcareservicesatenhancedratescomparedtousualMedicaidcosts.Newschaffer and colleagues conducted an evaluation of PCAP concentrating on the outcomesinsubstance-usingHIV-positivewomen.

ReichmanandFlorioreportanevaluationofNewJersey’sHealthStart program (New Jersey,USA).52 This enhanced antenatal care intervention was designed to increase the quantityandqualityofantenatalcare,withtheaimofincreasingbirthweightamongsocioeconomicallydisadvantagedwomen.TheprogrammedeliveredantenatalcaretoMedicaideligiblewomen,alongsideenhancedservicessuchascarecoordination.

“The key features of this program, available to pregnant Medicaid recipients, are an increased number of prenatal visits, increased provider reimbursement, case coordination with other social programs and integrated health support services such as psychological counselling and health education. Case managers, trained in cultural sensitivity, provide individualized plans of care and follow-up consultations through the pregnancy and for 60 days postpartum. To encourage women to get prenatal care early, community outreach efforts are mandated for all HealthStart providers. A system of presumptive eligibility, not part of the HealthStart program per se, was also established to enable financially eligible unenrolled pregnant women to obtain early care. The combination of provider supply incentives, enhanced services, and streamlined enrolment procedures was expected to increase the use of prenatal care and improve birth outcomes among Medicaid women in New Jersey.” 52

Effectiveness4.3

Outreach or other community-based interventions4.3.1

Theoverallstrengthandqualityofevidencerelatingtothesestudieswaspoor.Alloftheeleven studies evaluating the effect of outreach or other community-based interventions on the timing of initiation of antenatal care were observational study designs (nine cohortstudiesandtwobeforeandafterstudies),andonlyoneoftheelevenevaluationswas assessed as having adequate internal validity in relation to the outcome relevant to thisreview.Eightstudieswereassessedasclearlyreportingthetimingofinitiationofantenatalcareasanoutcomemeasure;intheremainingthreestudiesitwasunclearastowhetherthismeasurewasreportedasanoutcomemeasure.

Lay or paraprofessional home visiting and support

Rogers and colleagues assessed the effectiveness of a Resource Mothers intervention onthetimingofinitiationofantenatalcareamongpregnantteenagers,usingaretrospectiveobservationaldesign.Theevaluationusedtwodifferentcomparisongroups,onedrawnfromdifferentbutbroadlysimilargeographicalareas,andtheseconddrawnfrom adolescents who resided in the intervention areas before the intervention was implemented.Thestudywasconsideredtohavenomajorweaknesses,andwastheonlystudyincludedinthereviewtoadjustforpotentialconfoundingintheanalysisoftimingofinitiationofantenatalcare.Theevaluationreportsthatahigherproportionofinterventionadolescentsinitiatedantenatalcarebeforethefourthmonthofpregnancy(45%oftheinterventiongroupvs.41%inthegeographicalcomparatorgroupand40%inthe‘pre-intervention’comparatorgroup),withthisincreasesignificantincomparisontobothcontrolgroups.Anadjustedoddsratioforearlyinitiationofantenatalcareisreportedfortheinterventiongroupcomparedtothegeographicalcomparatorgroup(1.48,95%CI1.32,1.66)and‘pre-intervention’comparatorgroup(1.39,95%CI1.16,1.66).Theauthorsconcludedthatthestudydemonstratedastatisticallysignificantbeneficialeffectonthetimingofinitiationofantenatalcare.Becauseofpotentialforselectionbiaslargelyattributabletotheobservationalstudydesignandnon-randomassignmentofparticipants,thereviewersconsideredthestudyinconclusivebutconsistentwithapossiblebeneficialeffect.


Results are summarised in Table 4.

Alternative models of clinic-based antenatal care4.3.2

The quality of evidence relating to interventions involving alternative models of clinic-basedantenatalcarewasalsopoor.Fourofthefivestudiesinthiscategorywereobservationalcohortstudies,andonewasabeforeandafterstudywithoutacontemporaneouscomparatorgroup.Allfiveofthesestudieswereassessedashavingpoorinternalvalidityinrelationtotheoutcomerelevanttothisreview.However,thereviewers considered that in none of these studies was it clear whether timing of initiation ofantenatalcarewasreportedasanoutcomemeasure.Thisreflectedthefactthatmanyoftheseinterventionswereprimarilydesignedtoimproveantenatalcareutilisationasmeasuredbyattendanceforappointmentsratherthantimingofinitiation.

Results are summarised in Table 5.


Tab

le 4

. Eff

ect

iven

ess

of

ou

treach

or

oth

er

com

mu

nit

y-b

ase

d i

nte

rven

tio

ns

Au

tho

r an

d y

ear

Desi

gn

Ass

ess

men

t o

f in

tern

al

valid

ity

(1)

Tim

ing

of

an

ten

ata

l ca

re

init

iati

on

cl

earl

y a

n

ou

tco

me

measu

re?

Eff

ect

of

inte

rven

tio

n

on

tim

ing

of

init

iati

on

of

an

ten

ata

l ca

reR

evie

wer

com

men

ts

Au

tho

rs’

con

clu

sio

n (

2)i

Revie

wer

ass

ess

men

t (3

)

Cra

mer

2007

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Yes

XX

Nodatapresentedonbaselinecharacteristics,intervention

andcontrolgroupslikelytodifferwithrespecttoimportant

riskfactorsduetotargetingofintervention,noadjustment

forpotentialconfoundinginanalysisoftimingofinitiationof

ante

nat

al c

are.

Daa

lem

an

1997

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Yes

(+)

XNodatapresentedonbaselinecharacteristicsbyintervention/

controlstatus,interventionandcontrolgroupslikelyto

differwithrespecttoimportantriskfactorsduetotargeting

ofintervention,noadjustmentforpotentialconfoundingin

analysisoftimingofinitiationofantenatalcare,smallsample

size,outcomedataonlypresentedingraphicalform,authors’

conclusionsdonotappeartocorrespondtodatapresented.

Edger

ley

2007

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Yes

+X

Noadjustmentforpotentialconfoundinginanalysisoftiming

ofinitiationofantenatalcare,insufficientdatapresentedon

bas

elin

e ch

arac

terist

ics.

Hunte

2004

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

XX

Interventionandcontrolgroupsdifferwithrespectto

importantriskfactors,noadjustmentforpotentialconfounding

in a

nal

ysis

of

tim

ing o

f in

itia

tion o

f an

tenat

al c

are.

Jewell2000

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Yes

+X


ofinitiationofantenatalcare,inappropriatestatisticalanalysis

(matcheddata,unmatchedanalysis),unclearwhether

indiv

idual

s m

ay h

ave

bee

n r

efer

red b

y th

eir

ante

nat

al c

are

provider.

Julnes1994

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Yes

+X



inanalysisoftimingofinitiationofantenatalcare,small

samplesize.

Mac

kerr

as

2001

Bef

ore

and

afte

r st

udy

Poor

Yes

(+)

XNodatapresentedonbaselinecharacteristicsofcontrol(pre-

phase)andintervention(post-phase)groups,noadjustment

forpotentialconfoundinginanalysisoftimingofinitiationof

antenatalcare,noprotectionagainstsecularchanges.


Au

tho

r an

d y

ear

Desi

gn

Ass

ess

men

t o

f in

tern

al

valid

ity

(1)

Tim

ing

of

an

ten

ata

l ca

re

init

iati

on

cl

earl

y a

n

ou

tco

me

measu

re?

Eff

ect

of

inte

rven

tio

n

on

tim

ing

of

init

iati

on

of

an

ten

ata

l ca

reR

evie

wer

com

men

ts

Au

tho

rs’

con

clu

sio

n (

2)i

Revie

wer

ass

ess

men

t (3

)

Mas

on 1

990

Prospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

XX



in a

nal

ysis

of

tim

ing o

f in

itia

tion o

f an

tenat

al c

are.

Roger

s 1996

Retrospective

obse

rvat

ional

co

hort

st

udy

with

additio

nal

pre-

inte

rven

tion

comparator

group

Mix

edYe

s+

(+?)

Generallywelldesigned/reportedstudy,althoughsome

potentialforselectionbias.

Thompson

1998

Bef

ore

and

afte

r st

udy

Poor

Yes

XX


ofinitiationofantenatalcare,smallsamplesize,outcomedata

onlypresentedingraphicalform,noprotectionagainstsecular

chan

ges

.

Will

is 2

004

Prospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

XX

Interventionandcontrolgroupsdrawnfromdifferentsampling

frames,interventionandcontrolgroupsdifferwithrespectto


inanalysisoftimingofinitiationofantenatalcare,intervention

groupincludedincomparisongroup.

(1) Q

uali

ty a

ssess

men

t (G

ATE c

rite

ria)

(2) A

uth

ors

’ co

ncl

usi

on

(3) R

evie

wers

’ ass

ess

men

tGoodWellreportedandreliable;

MixedSomeweaknessesbutinsufficient

tohaveanimportanteffecton

usefulnessofstudy;

Poor

Studynotreliable,notuseful

+

StatisticallysignificantbeneficialeffectonPTB/IM

(+)

Effectconsistentwithbeneficialeffectbuteffectnotstatistically

significantand/orcautiousinterpretationoffindingsuggested

X

Noevidenceofbeneficialeffect

0

No c

oncl

usi

on s

tate

dN/ANotapplicable–outcomenotassessed

+

Studydemonstratesabeneficialeffect

(+?)

Stu

dy

inco

ncl

usi

ve b

ut

may

dem

onst

rate

abeneficialeffect

X

Studydoesnotprovideconvincing

evidenceofabeneficialeffect

N/ANotapplicable–outcomenotassessed

i Anumberoftheincludedstudiesreportedtimingofinitiationofantenatalcarebytrimesterorothersimilarmeasureandassessedwhethertherewasatrendtowardsearlierinitiation.The

conclusionreportedhererelatestothetestofeffectivenessreportedbytheauthor.


Tab

le 5

. Eff

ect

iven

ess

of

inte

rven

tio

ns

invo

lvin

g a

ltern

ati

ve m

od

els

of

clin

ic-b

ase

d a

nte

nata

l ca

re

Au

tho

r an

d y

ear

Desi

gn

Ass

ess

men

t o

f in

tern

al

valid

ity (

1)

Tim

ing

of

an

ten

ata

l ca

re

init

iati

on

cl

earl

y a

n

ou

tco

me

measu

re?

Tim

ing

of

an

ten

ata

l ca

re

init

iati

on

ou

tco

me r

esu

lts

Co

mm

en

ts

Au

tho

rs’

con

clu

sio

n (

2)

Revie

wer

ass

ess

men

t (3

)

Mar

tin 1

997

Bef

ore

and

afte

r st

udy

Poor

Uncl

ear

XX



inanalysisoftimingofinitiationofantenatalcare,small

samplesize,noprotectionagainstsecularchanges.

Morr

is 1

993

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

+X



in a

nal

ysis

of

tim

ing o

f in

itia

tion o

f an

tenat

al c

are.

Mvu

la 1

998

Prospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

XX



in a

nal

ysis

of

tim

ing o

f in

itia

tion o

f an

tenat

al c

are.

New

schaf

fer

1998

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

(+)

XNoadjustmentforpotentialconfoundinginanalysisoftiming

of

initia

tion o

f an

tenat

al c

are.

Rei

chm

an

1996

Retrospective

obse

rvat

ional

co

hort

stu

dy

Poor

Uncl

ear

0X



in a

nal

ysis

of

tim

ing o

f in

itia

tion o

f an

tenat

al c

are.

(1) Q

uali

ty a

ssess

men

t (G

ATE c

rite

ria)

(2) A

uth

ors

’ co

ncl

usi

on

(3) R

evie

wers

’ ass

ess

men

tGoodWellreportedandreliable;

MixedSomeweaknessesbutinsufficient

tohaveanimportanteffecton

usefulnessofstudy;

Poor

Studynotreliable,notuseful

+

StatisticallysignificantbeneficialeffectonPTB/IM

(+)

Effectconsistentwithbeneficialeffectbuteffectnotstatistically

significantand/orcautiousinterpretationoffindingsuggested

X

Noevidenceofbeneficialeffect

0

No c

oncl

usi

on s

tate

dN/ANotapplicable–outcomenotassessed

+

Studydemonstratesabeneficialeffect

(+?)

Stu

dy

inco

ncl

usi

ve b

ut

may

dem

onst

rate

abeneficialeffect

X

Studydoesnotprovideconvincing

evidenceofabeneficialeffect

N/ANotapplicable–outcomenotassessed

i Anumberoftheincludedstudiesreportedtimingofinitiationofantenatalcarebytrimesterorothersimilarmeasureandassessedwhethertherewasatrendtowardsearlierinitiation.The

conclusionreportedhererelatestothetestofeffectivenessreportedbytheauthor.


Discussion5

Principal findings5.1

Thepurposeofthisreviewwastosystematicallyidentifyandevaluatetheevidencerelatingtotheeffectivenessofinterventions,relevantinthecontextoftheNHS,whichaimtoincreasetheearlyinitiationofcomprehensiveantenatalcareinsociallydisadvantaged and vulnerable women.

Weidentifiedoverthirtypotentiallyrelevantinterventions(seeAnnexB)butonly16eligible evaluation studies: eleven of the sixteen related to community-based interventions involving outreach or community based services designed to increase the early initiation ofantenatalcare;andfivestudiesevaluatedtheeffectofalternativemodelsoforganisingand delivering antenatal care on the timing of initiation of antenatal care.

Oftheelevenstudiesrelatingtocommunity-basedinterventions,threeevaluatedinterventionswhichconsistedsolelyofsocialsupportand/orhomevisitsdeliveredbylayorparaprofessionalworkers,oneevaluatedtheprovisionofbilingual‘linkworkers’(workinginbothprimarycareandobstetricclinics),oneevaluateda‘mobilewomen’shealthbus’,andsixevaluatedother,multi-componentinterventions.

Ofthefivestudiesrelatingtoalternativemodelsoforganisinganddeliveringantenatalcare,twoevaluated‘teenclinics’,oneevaluateda‘neighbourhoodclinic’andtwoevaluated‘enhancedprenatalcare’models.

We found eligible studies relating to only a few of the disadvantaged and vulnerable groupsofinterest:fourinterventionstargetedpregnantteenagers,seventargetedand/orwereevaluatedinsociallydisadvantaged‘ethnicminority’populations(includingAustralianindigenouswomenandnon-nativelanguagespeakers),andfivewereaimedatsocio-economicallydisadvantagedwomen.Wedidnotfindeligiblestudiesrelatingtointerventions seeking to increase early initiation of antenatal care in other vulnerable or ‘atrisk’subgroupssuchashomelesswomen,travellers,refugees,substanceandalcoholusersandwomenwithmentalhealthproblemsorlearningdisabilities.

Overall,thequalityofevidencewaspoor.Wedidnotidentifyanyeligiblerandomisedcontrolledtrials(RCTs)andonlyonestudy-aretrospectivecohortstudywithanadditionalpre-interventioncomparatorgroup-wasassessedashavingadequateinternalvalidity.45 This study evaluated a Resource Mothers Program,whichusedparaprofessionalwomentodeliversocialsupport,healthpromotion/educationandotherassistancetopregnantadolescentsathomeandforoneyearafterdelivery.Theevaluation,whichwasconductedinapredominantlyblack,non-urbanUSpopulationfoundthattheinterventionwaseffectiveinincreasingtheproportionofpregnantadolescentsinitiatingantenatalcarebythefourthmonthofpregnancy.ThisinterventioncouldbeconsideredtoaddressbarrierstocaregroupedundertwoofthethreeheadingssuggestedbyCooper’saccessmodel.21Personalandfamilybarrierswereaddressedbytheprovisionofculturally-appropriateantenataleducationandsocialsupport,deliveredbythe‘resourcemothers’,many of whom had been teenage mothers themselves. The resource mothers facilitated access to antenatal care by acting as an advocate and drawing attention to the needs of the adolescents within the healthcare system. Structural barriers to care were attenuated bytheresourcemotherfollowingupappointmentsandarrangingtransportation.Therefore,thisinterventionmovedbeyondsimplyprovidingservices,anapproachcriticisedbyLavenderandcolleagues,10whilealsotakingintoaccountthecomplexinterplaybetweenindividualsandhealthcareservices.27,57ItwouldalsoappeartoaddressthedifferentialconceptualisationofhealthdescribedbyDixon-Woodsasassociatedwithsociallydisadvantagedgroups,27inparticularthelackofappreciationofpreventivecare,asoneoftherolesoftheresourcemotherwastoemphasisetheneedfor“earlyandregularprenatalcare”.TheinterventionwasadequatelydescribedandcontainedsomepotentiallytransferableelementsbutthegeneralisabilityofthefindingstoaUKpopulationisunknown.Forexample,routesofreferralintotheprogrammeincludedWIC(aUSspecificwelfareprogramme)andchurches.


Evidence relating to other interventions was inconclusive due to the methodological limitationsoftheincludedstudies.However,althoughtheireffectivenessisunproven,someoftheinterventionsidentifiedinthisreviewincludedelementsofpotentialrelevanceintheUKwhichthereviewersconsideredmightplausiblyaffectthetimingofinitiationofantenatalcareinsociallydisadvantagedandvulnerablegroups.Theseincluded:

Mobilehealthclinics,providingfreewalk-inorappointmentservicesincludinginitial•antenatal care.49 This intervention strategy may address structural barriers to care suchaslackoftransportationortheneedtonegotiateanappointmentsystem.Another outcome evaluation of a similar intervention in the USA has recently been published58(outsidethetimeperiodofthisreview),reportingabeneficialeffectoftheinterventionconsistentwiththeconclusionsoftheevaluationincludedinthepresentreview.

LinkworkerssituatedinGPsurgeries,actingas“‘facilitators’and‘interpreters’whilst•alsofulfillinganeducativerole”.42 This form of intervention may work well for some ethnicminoritygroupsandwomenforwhomlanguagedifficultiesmaybeabarriertoantenatalcare.Theincludedevaluationshowednoeffectontheproportionofwomenbookingbefore12weeks.However,theincludedstudywasnotwelldesignedtoevaluate the effect on this outcome.

Culturallyappropriatecommunity-basedprogrammes,wherelaywomenencourage•greateruseofantenatalcarethroughintegratingtraditionalbeliefsandpracticesalongside more conventional antenatal education.41 Programmes such as this are most likelytoinfluencepersonalbarrierstocaresuchasacceptability,attitudes/beliefsandculturalpreferences.AlthoughthisinterventiontargetsasubgroupwhichhasnodirectlyequivalentgroupintheUK,theemphasisonaddressingculturalbeliefsandpracticesisconsideredrelevanttoethnicminoritygroupsintheUK.

Theseinterventionsmeritfurtherconsiderationandpossiblyfurther,morerobustevaluationinaUKsetting.

Strengths and limitations of this systematic review5.2

Weusedacomprehensive,multi-stagesearchstrategywhichenabledustoidentifyawiderange of relevant interventions described in the literature. The relatively small number of studieseligibleforinclusioninthissystematicreviewreflectsthepaucityofeffectivenessevaluations in this area.

Wedidnotrestrictinclusiontospecificstudydesigns,otherthanrequiringsomeformofcomparator/controlgroup,andhencethematerialdescribedherereflectsthebreadthoftheeffectivenessevidenceavailableinthescientificliterature.Givensomeofthereportinglimitationsoftheincludedmaterial,wefounditchallengingtodevelopreproducibleinclusion/exclusioncriteriarelatingtotheaimsoftheintervention/evaluation.Weresolvedthis by including studies where the reviewers could not easily reach a consensus as to whetherornotthetimingofinitiationofantenatalcarewasreportedasanoutcomemeasure.Weconsideredthatthisinclusiveapproachwaspreferabletoexcludingpotentiallyrelevantstudiesbutaconsequenceisthatwehaveincludedsomestudies–particularlythoserelatingtoalternativemodelsoforganisinganddeliveringclinic-basedantenatal care – of questionable relevance. A further consequence of this was that we assessed internal validity of the study in relation to the estimated effect of the intervention onthetimingofinitiationofantenatalcare,evenwhenthiswasnotnecessarilytheaimofthestudy.Thisenabledustoassesswhetherthestudyprovidedrobustevidenceofaneffectontimingofinitiationofantenatalcare.However,ourqualityassessmentsshouldnotbeinterpretedasreflectingthequalityofthestudyinrelationtotheaimsstatedbythe author where these are different from the effectiveness question addressed by this review.

Forpragmaticreasons,wedidnotincludeevaluationsreportedinthegreyliterature.Weidentifiedbutdidnotincludeasmallnumberofpotentiallyrelevantstudiesinthegreyliterature:thesewerepredominantlyidentifiedthroughscreeningreferencesof


includedstudies,ratherthanthroughthebibliographicdatabasesearches,andmostrelated to interventions targeting Australian indigenous women. Although the inclusion ofsuchstudieswouldpotentiallyhaveaddedtothedescriptiveelementsofthisreview,weconsideritunlikelythattheinclusionofsuchreportswouldhaveinfluencedourconclusions regarding effectiveness.

Thegeneralisabilityoffindingsandthetransferabilityofinterventionspresentamethodologicalchallengeinreviewsofthiskind.Bydevelopinginclusioncriteriabasedonaconceptual‘barrierstocare’model21wewereabletooperationalisecriteriathatenabledustoexcludeanumberofstructuralandfinancialinterventionsnotrelevantinthecontextofapubliclyfundeduniversalhealthcaresystem.Wewerethusabletofocuson interventions most likely to be relevant in the context of the NHS. We note above some issuesrelatingtobothtransferabilityandgeneralisabilityoffindingstootherpopulationsbutthisisanareawherefurthertheoreticalworktodevelopaconceptualframeworkmightbehelpful.

Findings in relation to other published evidence5.3

Threepublishedliteraturereviewshaveevaluatedtheeffectofdifferentantenatalintervention strategies on a range of outcomes and have included results relating to the timing of initiation of antenatal care. Rumbold and Cunningham evaluated the effect of changes in the delivery of antenatal care on outcomes for Australian indigenous women.30Fouroftheteninterventionsincludedinthisreviewreportedtimingofinitiationof antenatal care as an outcome.59–62Twoofthesereportedastatisticallysignificantbeneficialeffectontimingofinitiationofantenatalcare:onewasacommunitybasedsupportprogrammeforpregnantwomen(includedinthepresentreview),60 and the other a“culturallyappropriatemidwiferyprogram”.59 The latter intervention was not included inthepresentreviewastherelevantresultswerereportedonlyinthegreyliterature.NeitherofthetwointerventionsidentifiedintheRumboldreviewashavingnoeffectontimingofinitiationwereincludedinthepresentreview,bothbecausethecomparatorgroupsdidnotmeetoureligibilitycriteria.62,63Indiscussingtheirresults,theauthorscommentonthechallengesofsynthesisingresultsacrossdifferentstudies,referringinparticulartothelackofconsistencyinoutcomesandthediversityofcomparisongroups.However,theyconcludethattheresultssuggest“modestincreasesinindicatorsofantenatalcareutilization,mostnotablyincreasesintheproportionofwomenaccessingantenatalcareinthefirsttrimester”.30

A second relevant literature review by Persily evaluated the effect of lay home visiting onpregnancyoutcomes.31InPersily’sreview,alleightstudiesthatreportedaneffectonantenatalcareusefoundabeneficialeffectoftheintervention.However,onlyfourofthesestudieslookedspecificallyattimingofinitiationofantenatalcare.Threeofthesestudiesareincludedinthepresentreviewanddescribedinsomedetailinearliersections.45,48,55ThefourthstudyevaluatedtheeffectofalayhomevisitingprogrammetargetedatHispanicpregnantwomeninanurbanarea,64andwasexcludedinthepresentreviewbecauseitusedanineligiblecontrolgroup.Theauthorofthisreviewhighlightedthemethodologicalweaknessesofincludedstudies,butneverthelessconcludedthat“layworkersmaybeespeciallysuccessfulin…impactingonsocialandenvironmentalriskfactors as well as on health care utilization”.31

The review conducted by D’Souza and Garcia32 considered a variety of different interventionstoimproveperinataloutcomes,evaluatedindifferentsubgroupsofdisadvantagedwomen.Onlyoneinterventiondescribedintheirreportlookedattimingof initiation of antenatal care. This intervention - health advocacy for ethnic minority women65-wasassessedasunlikelytohaveabeneficialimpactonlatebookingforantenatalcare.Thisinterventionwasnotincludedinthepresentreviewbecausethetimingofinitiationofantenatalcarewasreportedasacontinuousmeasure(meangestationalageat‘booking’),anditwasnotpossibletoderivetheproportionofwomenbookingbyagivendatefromthedatareported.D’SouzaandGarciacommentonthe


limitedevidenceofeffectivenessacrossallthestudiesreviewed,concludingthatlittleornoreliableevidenceisavailableregardingpromisinginterventionsapplicabletodisadvantagedgroupsofpregnantwomenintheUK.32

Overall,ourfindingsarebroadlyconsistentwiththeresultsofthesepreviousreviews,and echo the authors’ conclusions about the methodological limitations of the available evidence.

Implications and recommendations5.4

Theresultsofthisreview,consideredalongsidetheexistingliteratureonthistopic,suggestthatthereisinsufficientevidencetorecommendthatanyoftheinterventionsdescribedintheliteratureshouldbeimplementedasameansofincreasingtheearlyinitiationofantenatalcareinsociallydisadvantagedandvulnerablegroupsofpregnantwomen.Inreviewingtheincludedstudies,wefocussedspecificallyonthetimingofinitiationofantenatalcare,althoughthiswasnottheprimaryoutcomeofmanyoftheincludedstudies.Ourfindingsdonotthereforenecessarilyindicatethattheincludedinterventionsdonothaveabeneficialeffectonotheroutcomes,forexampleimprovedadherencetoarecommendedscheduleofappointmentsonceantenatalcareisinitiated.

Majormethodologicaland/orreportingweaknesseswereidentifiedinmanyoftheincludedstudies.Suchfundamentalmethodologicalflawsneedtobeavoidedthroughthecarefuldesignoffutureevaluations.Poorreporting,anotherweaknessobservedinmanyofthestudies,mightbeminimisedbyadherencetorelevantreportingguidelinessuch as SQUIRE66(qualityimprovementstudies)andSTROBE67 (observational studies in epidemiology).

Alloftheevaluationsincludedinthisreviewwereobservationalstudies.Thepotentialweaknesses of such study designs have been well documented. RCTs are considered themostrobustdesignforassessingeffectiveness,althoughweareawarethatmanyof the interventions considered in this review would have been challenging to evaluate usingstandardrandomisedapproachessuchasclusterorindividuallyrandomisedRCTs.However,avarietyofexperimentalmethodspotentiallysuitablefor‘complex’interventionshavebeenproposed.68Furthermore,itmaybepossibletogreatlyimprovethequalityofevaluationswithoutrecoursetostandardrandomiseddesigns.Forexampleacontrolledbeforeandafterstudy(CBA)canprovidemoderatelyrobustevidenceprovidedthatthestudyiscarefullyplannedandconductedandthecontrolgroupisappropriatelyselectedtocreatestudygroupswithsimilar‘baseline’characteristics.

Although we did not identify interventions for which there was sound evidence of effectiveness,ourreviewneverthelessidentifiedanumberofinterventionsthatcouldplausiblyaffectthetimingofinitiationofantenatalcareandwhichwereconsideredtobepotentiallyrelevantintheUKcontext.Someofthesemightprovideameansofaddressingthe concern raised by Dixon-Woods et al. in their review on access to healthcare for vulnerablegroupsthat“manyinterventionsandpoliciesarenotwellmatchedtowhatwehaveidentifiedasthemajorbarrierstoaccess”.24Thematerialidentifiedduringthis review (including the interventions described in the literature for which no eligible evaluationswerefound)providesasourceofdatathatmightbefurther‘mined’toidentifytheinterventionswhichmostplausiblyaddressthebarrierstoaccessingantenatalcareexperiencedbysociallydisadvantagedandvulnerablegroupsintheUK.Inparticular,furtherworkmightusefullybeundertakentoexploreanddescribethemechanismsofactionandbarriersaddressedbysomeofthemorerelevantinterventions,incombinationwith a synthesis of the qualitative literature aimed at identifying the barriers to and facilitatorsofantenatalcareuptakebysociallydisadvantagedandvulnerablesubgroupsintheUK.Asynthesisofthesetwosetsoffindingscouldpotentiallyguidefutureservicedevelopmentandresearchprioritiesbyidentifyingtheinterventionswhichbestaddressthe‘barriersandfacilitators’relevantintheUKcontext.


Conclusion5.5

Insummary,wefoundinsufficientevidencetoconcludethatinterventionsthataimtoincrease the early initiation of antenatal care in socially disadvantaged and vulnerable populationsofwomenareeffective.However,theabsenceofevidenceshouldnotbeinterpretedasevidencethattheinterventionsevaluatedarenecessarilyineffective.Oneinterventionbasedonhomevisitingforpregnantadolescentswasconsidered‘promising’,and several other intervention strategies were considered to contain elements that would meritfurtherconsiderationandpossiblyevaluation.Overall,theresultsofthisreviewhighlightthepaucityofevidenceandtheneedforfurtherwelldesignedevaluationsto ensure that services designed to increase the early initiation of antenatal care are evidence based.

AcknowledgementThisisanindependentreportfromastudywhichisfundedbythePolicyResearchProgrammeintheDepartmentofHealth.TheviewsexpressedarenotnecessarilythoseoftheDepartment.


ReferencesNational Institute for Clinical Excellence (NICE). Antenatal care: routine care for the 1. healthypregnantwoman(clinicalguideline).London:NICE,2008.

ConfidentialEnquiryintoMaternalandChildHealth(CEMACH).2. Perinatal Mortality 2007.London:CEMACH,2009.

DepartmentofHealth.3. Implementation Plan for Reducing Health Inequalities in Infant Mortality: A Good Practice Guide.London:DH,2007.

House of Commons Health Committee. 4. Inequalities in access to maternity services: eighth report of session 2002‑03.London:TheStationeryOffice,2003.http://www.parliament.the‑stationery‑office.co.uk/pa/cm200203/cmselect/cmhealth/696/696.pdf

DepartmentofHealth.5. Maternity Matters: Choice, access and continuity of care in a safe service.London:DH,2007.

DepartmentofHealth.6. The NHS in England: the Operating Framework for 2008/09. London:DH,2007.

HM Government. 7. PSA Delivery Agreement 19: Ensure better care for all. London: HMSO,2007.

RedshawM,RoweRE,HockleyC,BrocklehurstP.8. Recorded delivery: a national survey of women’s experience of maternity care. Oxford: National Perinatal EpidemiolgyUnit,2006.

RoweRE,GarciaJ.Socialclass,ethnicityandattendanceforantenatalcareinthe9. UnitedKingdom:asystematicreview.J Public Health2003;25(2):113-119.

LavenderT,DowneS,FinnlaysonK,WalshD.10. Access to antenatal care: a systematic review:UniversityofCentralLancashire,2007.

KupekE,PetrouS,VauseS,MareshM.Clinical,providerandsociodemographic11. predictorsoflateinitiationofantenatalcareinEnglandandWales.BJOG: An International Journal of Obstetrics and Gynaecology2002;109(3):265-273.

RoweRE,MageeH,QuigleyMA,HeronP,AskhamJ,BrocklehurstP.Socialand12. ethnic differences in attendance for antenatal care in England. Public Health2008;122(12):1363-1372.

EssexC,CounsellAM,GeddisDC.Thedemographiccharacteristicsofearlyand13. late attenders for antenatal care. Australian & New Zealand Journal of Obstetrics & Gynaecology1992;32(4):306-308.

DelvauxT,BuekensP,GodinI,BoutsenM.BarrierstoprenatalcareinEurope.14. American Journal of Preventive Medicine2001;21(1):52-59.

TrinhL,RubinG.LateentrytoantenatalcareinNewSouthWales,Australia.15. Reproductive Health2006;3(1):8.

AlderliestenME,VrijkotteTGM,vanderWalMF,BonselGJ.Latestartofantenatal16. careamongethnicminoritiesinalargecohortofpregnantwomen.BJOG: An International Journal of Obstetrics & Gynaecology2007;114(10):1232-1239.

RaatikainenK,HeiskanenN,HeinonenS.Under-attendingfreeantenatalcareis17. associatedwithadversepregnancyoutcomes.BMC Public Health2007;7(1):268.

LalchandaniS,MacQuillanK,SheilO.Obstetricprofilesandpregnancyoutcomesof18. immigrant women with refugee status. Irish Medical Journal2001;94(3):79-80.

http://www.parliament.the-stationery-office.co.uk/pa/cm200203/cmselect/cmhealth/696/696.pdf




BravemanP,MarchiK,EgerterS,PearlM,NeuhausJ.BarrierstoTimelyPrenatal19. CareAmongWomenWithInsurance:TheImportanceofPrepregnancyFactors.Obstetrics & Gynecology2000;95(6,Part1):874-880.

Gulzar L. Access to Health Care. 20. Journal of Nursing Scholarship1999;31(1):13-19.

CooperLA.Designingandevaluatinginterventionstoeliminateracialandethnic21. disparitiesinhealthcare.Journal of General Internal Medicine2002;17(6):477.

SzczepuraA.Accesstohealthcareforethnicminoritypopulations.22. Postgrad Med J 2005;81(953):141-147.

BraggR,CookJ.MoneyandMaternity:chargingvulnerablepregnantwomenfor23. NHS care. UK Public Health Association Conference.Brighton,UK,2009.

DixonWoodsM,KirkD,AgarwalS,AnnandaleE,ArthurT,HarveyJ,etal.24. Vulnerable groups and access to health care:a critical interpretive review: NHS Service Delivery andOrganisationR&D(NCCSDO),2005.

NealeJ.Barrierstoaccessinggenerichealthandsocialcareservices:aqualitative25. study of injecting drug users. Health & Social Care in the Community2008;16(2):147.

DowneS,FinlaysonK,WalshD,LavenderT.‘Weighingupandbalancingout’:a26. meta-synthesis of barriers to antenatal care for marginalised women in high-income countries. BJOG: An International Journal of Obstetrics & Gynaecology2009;116(4):518-529.

Dixon-WoodsM,CaversD,AgarwalS,AnnandaleE,ArthurA,HarveyJ,etal.27. Conductingacriticalinterpretivesynthesisoftheliteratureonaccesstohealthcarebyvulnerablegroups.BMC Medical Research Methodology2006;6:35.

AchesonD.IndependentInquiryintoInequalitiesinHealth.London:HMSO,1998.28.

ShortR.Perinatalandneonatalmortality.SecondReportfromtheSocialServices29. Committee1979-80.London:HMSO,1980.

RumboldA,CunninghamJ.AReviewoftheImpactofAntenatalCareforAustralian30. IndigenousWomenandAttemptstoStrengthentheseServices.MaternalandChildHealthJournal2008;12(1):83-100.

PersilyCA.LayHomeVisitingMayImprovePregnancyOutcomes.HolisticNursing31. Practice2003;17(5):231-238.

D’SouzaL,GarciaJ.Improvingservicesfordisadvantagedchildbearingwomen.32. Child:Care,Health&Development2004;30(6):599-611.

FiscellaK.Doesprenatalcareimprovebirthoutcomes?Acriticalreview.Obstetrics&33. Gynecology1995;85(3):468-479.

BryceRL,StanleyFJ,GarnerJB.Randomizedcontrolledtrialofantenatalsocial34. supporttopreventpretermbirth.BritishJournalofObstetricsandGynaecology1991;98(10):1001-1008.

MooreML,MeisPJ,ErnestJM,WellsHB,ZaccaroDJ,TerrellT.Arandomizedtrial35. ofnurseinterventiontoreducepretermandlowbirthweightbirths.ObstetricsandGynecology1998;91(5Pt1):656-661.

KotelchuckM.AnevaluationoftheKessnerAdequacyofPrenatalCareIndexand36. aproposedAdequacyofPrenatalCareUtilizationIndex.AmericanJournalofPublicHealth1994;84(9):1414-1420.


OrganisationforEconomicCo-operationandDevelopment(OECD).OECDHealth37. Data 2009: Statistics and Indicators for 30 Countries: http://www.oecd.org/health/healthdata, 2009.

HollowellJ,GrayR,KurinczukJJ,OakleyL,BrocklehurstP.Asystematicreviewof38. theeffectivenessofantenatalcareprogrammestoreduceinfantmortalityanditsmajor causes in socially disadvantaged and vulnerable women Oxford: National PerinatalEpidemiologyUnit,2009.

JacksonR.TheGATEframe:criticalappraisalwithpictures.Quality&Safetyin39. HealthCare2006;11(2):35.

ThomasJ.Informationsystemsforconductingsystematicreviews:acasestudy.40. SixthInternationalCampbellColloquium.LosAngeles,22-24February2006.Posteravailable at http://eppi.ioe.ac.uk/cms/.

MackerrasD.BirthweightchangesinthepilotphaseoftheStrongWomenStrong41. BabiesStrongCultureProgramintheNorthernTerritory.Australian&NewZealandJournalofPublicHealth2001;25(1):34-40.

MasonES.TheAsianMotherandBabyCampaign(theLeicestershireexperience).42. JournaloftheRoyalSocietyofHealth1990;110(1):1-4.

ThompsonM,CurryMA,BurtonD.Theeffectsofnursingcasemanagementonthe43. utilizationofprenatalcarebyMexican-AmericansinruralOregon.PublicHealthNursing1998;15(2):82-90.

MartinRD,MacDowellNM,MacmannJM.Effectivenessofateenpregnancyclinicina44. managedcaresetting.ManagedCareQuarterly1997;5(3):20-7.

RogersMM,Peoples-ShepsMD,SuchindranC.Impactofasocialsupportprogram45. onteenageprenatalcareuseandpregnancyoutcomes.JournalofAdolescentHealth1996;19(2):132-40.

JewellNA,RussellKM.Increasingaccesstoprenatalcare:anevaluationofminority46. healthcoalitions’earlypregnancyproject.JournalofCommunityHealthNursing2000;17(2):93-105.

CramerME,ChenL-W,RobertsS,CluteD.Evaluatingthesocialandeconomicimpact47. ofcommunity-basedprenatalcare.PublicHealthNursing2007;24(4):329-36.

DaalemanTP.Theeffectofaparaprofessionalhomevisitingprogramonutilizationof48. prenatalcare.KansasMedicine1997;98(2):6-9.

EdgerleyL,El-SayedY,DruzinM,KiernanM,DanielsK.UseofaCommunityMobile49. HealthVantoIncreaseEarlyAccesstoPrenatalCare.MaternalandChildHealthJournal2007;11(3):235-239.

HunteHER,TurnerTM,PollackHA,LewisEY.Abirthrecordsanalysisofthematernal50. infanthealthadvocateserviceprogram:AparaprofessionalinterventionaimedataddressinginfantmortalityinAfricanAmericans.EthnicityandDisease2004;14(3SUPPL. 1):S1-102-S1-107.

MorrisDL,BerensonAB,LawsonJ,WiemannCM.Comparisonofadolescent51. pregnancyoutcomesbyprenatalcaresource.JournalofReproductiveMedicinefortheObstetricianandGynecologist1993;38(5):375-380.

ReichmanNE,FlorioMJ.TheeffectsofenrichedprenatalcareservicesonMedicaid52. birthoutcomesinNewJersey.JournalofHealthEconomics1996;15(4):455-76.

http://www.oecd.org/health/healthdata, 2009

http://www.oecd.org/health/healthdata, 2009

http://eppi.ioe.ac.uk/cms/


WillisWO,EderCH,LindsaySP,ChavezG,SheltonST.LowerRatesofLow53. Birthweight and Preterm Births in the California Black Infant Health Programs. JournaloftheNationalMedicalAssociation2004;96(3):315-324.

MvulaMM,MillerJrJM.Acomparativeevaluationofcollaborativeprenatalcare.54. ObstetricsandGynecology1998;91(2):169-173.

JulnesG,KonefalM,PindurW,KimP.Community-basedperinatalcarefor55. disadvantagedadolescents:evaluationofTheResourceMothersProgram.JournalofCommunityHealth1994;19(1):41-53.

NewschafferCJ,CocroftJ,HauckWW,FanningT,TurnerBJ.Improvedbirth56. outcomesassociatedwithenhancedMedicaidprenatalcareindrug-usingwomeninfectedwiththehumanimmunodeficiencyvirus.ObstetricsandGynecology1998;91(6):885-891.

SwordW.Asocio-ecologicalapproachtounderstandingbarrierstoprenatalcarefor57. womenoflowincome.JournalofAdvancedNursing1999;29(5):1170-1177.

O’ConnellE,ZhangG,LeguenF,PrinceJImpactofaMobileVanonPrenatalCare58. Utilization and Birth Outcomes in Miami-Dade Country. Maternal and Child Health Journal2009,DOI:10.1007/s10995-009-0496-8(publishedonline:15August2009)

Daruk Aboriginal Community Controlled Medical Service. Evaluation of the Daruk 59. Antenatal Program 1998. Sydney: Daruk Aboriginal Medical Service and Western PublicHealthUnit,1998.

MackerrasD.EvaluationoftheStrongWomen,StrongBabies,StrongCulture60. Program:Resultsfortheperiod1990–1996inthethreepilotcommunities.IssueNo.2/98.Darwin,NorthernTerritory:MenziesOccasionalPapers,1998.

PerkinsS,SimonS.TheNguaGundi(Mother/Child)Project1997–2000.Queensland,61. Australia:QueenslandHealth,2000.

CarterE,LumleyJ,WilsonG,BellS.‘Alukura...formydaughtersandtheirdaughters62. andtheirdaughters’.AreviewofCongressAlukura.AustralianandNewZealandJournalofPublicHealth2004;28(3):229-234.

DormanR.TheNguaGundi(motherandchild)program.AboriginalandIslander63. HealthWorkerJournal1997;21(5):2-6.

BrayM,EdwardsL.AprimaryhealthcareapproachusingHispanicoutreachworkers64. asnurseextenders.PublicHealthNursing1994;11(1):7-11.

ParsonsL,DayS.Improvingobstetricoutcomesinethnicminorities:anevaluation65. ofhealthadvocacyinHackney.JournalofPublicHealth1992;14(2):183-191.

OgrincG,MooneySE,EstradaC,FosterT,GoldmannD,HallLW,etal.TheSQUIRE66. (StandardsforQUalityImprovementReportingExcellence)guidelinesforqualityimprovementreporting:explanationandelaboration.QualSafHealthCare2008;17(Suppl_1):i13-32.

VandenbrouckeJP,vonElmE,AltmanDG,GøtzschePC,MulrowCD,PocockSJ,et67. al.StrengtheningtheReportingofObservationalStudiesinEpidemiology(STROBE):ExplanationandElaboration.PLoSMed2007;4(10):e297.

CraigP,DieppeP,MacintyreS,MichieS,NazarethI,PetticrewM.Developingand68. evaluatingcomplexinterventions:thenewMedicalResearchCouncilguidance.BMJ2008;337:a1655.


MurrellNL,ScherzerT,RyanM,FrappierN,AbramsA,RobertsC.TheAfterCare69. Project: an intervention for homeless childbearing families. Family & Community Health2000;23(3):17-27.

CovingtonD,Peoples-ShepMD,PaulA.BuescherPA,BennettTA,PaulMV.An70. EvaluationofanAdolescentPrenatalEducationProgram.AmericanJournalofHealthBehavior1998;22(5):323-333.

JanssonLM,SvikisD,LeeJ,PaluzziP,RutiglianoP,HackermanF.Pregnancyand71. addiction.Acomprehensivecaremodel.JournalofSubstanceAbuseTreatment1996;13(4):321-9.

AffonsoDD,MayberryLJ,GrahamK,ShibuyaJ,KunimotoJ.PrenatalandPospartum72. CareinHawaii:ACommunity-BasedApproach.JournalofObstetric,Gynecologic,&NeonatalNursing1993;22(4):320-328.

JanS,ConatyS,HeckerR,BartlettM,DelaneyS,CaponT.Anholisticeconomic73. evaluationofanAboriginalcommunity-controlledmidwiferyprogrammeinWesternSydney.JHealthServResPolicy2004;9(1):14-21.

McFarlaneJ.DeMadresaMadres:anaccessmodelforprimarycare.American74. JournalofPublicHealth1996;86(6):879-80.

McFarlaneJ,FehirJ.DeMadresaMadres:Acommunity,primaryhealthcare75. programbasedonempowerment.HealthEducationQuarterly1994;21(3):381-394.

McFarlaneJ,KellyE,RodriguezR,FehirJ.DeMadresaMadres:womenbuilding76. communitycoalitionsforhealth.HealthCareforWomenInternational1994;15(5):465-76.

RodriguezR,McFarlaneJ,MahonJ,FehirJ.DeMadresaMadres:acommunity77. partnershiptoincreaseaccesstoprenatalcare.BulletinofthePanAmericanHealthOrganization1993;27(4):403-8.

JeffersDF.Outreachchildbirtheducationclassesforlow-incomefamilies:astrategy78. forprogramdevelopment.AWHONNSClinicalIssuesinPerinatal&WomensHealthNursing1993;4(1):95-101.

OvreboB,RyanM,JacksonK,HutchinsonK.TheHomelessPrenatalProgram:A79. modelforempoweringhomelesspregnantwomen.HealthEducationQuarterly1994;21(2):187-198.

ClarkeLL,MillerMK,VogelWB,DavisKE,MahanCS.TheeffectivenessofFlorida’s80. “ImprovedPregnancyOutcome”program.JournalofHealthCareforthePoor&Underserved1993;4(2):117-32.

ClintonB.TheMaternalInfantHealthOutreachWorkerproject:Appalachian81. communitieshelptheirown.In:LarnerM,HalpernR,HarkavyO,editors.FairStartforChildren:Lessonslearnedfromsevendemonstrationprojects.NewHaven:YaleUniversityPress,1992:23-45.

TessaroI,CampbellM,O’MearaC,HerrickH,BuescherP,MeyerR,etal.Statehealth82. departmentanduniversityevaluationofNorthCarolina’sMaternalOutreachWorkerProgram.AmericanJournalofPreventiveMedicine1997;13((6Suppl)):38-44.

PanarettoKS,LeeHM,MitchellMR,LarkinsSL,ManessisV,BuettnerPG,etal.83. ImpactofacollaborativesharedantenatalcareprogramforurbanIndigenouswomen:aprospectivecohortstudy.MedicalJournalofAustralia2005;182(10):514-9.


AndrusNC,PartnerSF,LeppertPC.Analyzingstrategiesfordevelopingaprenatal84. healthcareoutreachprogramtoreducesocialandculturalbarriers.JournalofHealth&HumanServicesAdministration1997;20(2):230-41.

DwyerC,CookeH,HortK.Bringingaccessibleantenatalcaretodisadvantaged85. womeninouterwesternSydney.AustralianJournalofPrimaryHealth2004;10(3):67-75.

ClarkM.ThePeerSupportProgramme:youngmothersofferingsupporttoyoung86. pregnantwomeninaclinicsetting.BirthIssues2000;9(2):47-53.

RegueroW,CraneM.ProjectMotherCare:onehospital’sresponsetothehigh87. perinataldeathrateinNewHaven,CT.PublicHealthReports1994;109(5):647-52.

NagyM,LeeperJD,Hullet-RobertsonS,NorthrupRS.TheRuralAlabama88. PregnancyandInfantHealthproject:Aruralclinicreachesout.In:LarnerM,HalpernR,HarkavyO,editors.FairStartforChildren:Lessonslearnedfromsevendemonstrationprojects.NewHaven:YaleUniversityPress,1992:91-114.

BaldwinM,StevensonY.Amodelforprovidingprenatalhealthcaretoindigenous89. womenlivinginremoteareas.InternationalJournalofCircumpolarHealth2001;60(4):623-31.

MattsonS,LewL.CulturallysensitiveprenatalcareforSoutheastAsians.JOGNN-90. JournalofObstetric,Gynecologic,&NeonatalNursing1992;21(1):48-54.

FlynnL,BuddM,ModelskiJ.EnhancingResourceUtilizationAmongPregnant91. Adolescents.PublicHealthNursing2008;25:140-148.

ReeceEA,LeguizamonG,SilvaJ,WhitemanV,SmithD.Intensiveinterventional92. maternitycarereducesinfantmorbidityandhospitalcosts.JournalofMaternal-Fetal&NeonatalMedicine2002;11(3):204-10.

DewartT,ZaengleD.TheDoor’sPerinatalProgramforPregnantandParentingTeens.93. JournalofPerinatalEducation2000;9(2):39-46.

MeisterJS,WarrickLH,deZapiénJG,WoodAH.Usinglayhealthworkers:Case94. studyofacommunity-basedprenatalintervention.JournalofCommunityHealth1992;17(1):37-51.


Annex A: Medline search strategyDatabase: Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations and Ovid MEDLINE(R) <1950 to Present>

Search Strategy:

--------------------------------------------------------------------------------

1expSocioeconomicFactors/orexpSocialClass/

2(equityorinequalit$orequalit$orunequal$orinequit$ordisparit$orgaporgapsorgradient$ordisadvantag$orsocioeconomic$).ti,ab.

3healthinequalit$.mp.orHealthStatusIndicators/or*HealthStatusDisparities/or*HealthcareDisparities/

4expPoverty/orexpMedicalIndigency/orvulnerablepopulations/

5expMinorityHealth/orexpMinorityGroups/orpopulationgroups/orexpethnicgroups/orhealthservices,indigenous/

6(ethnicor(blackadj2asian)).ti,ab.

7(multiethnic$ormultiethnic$ormultiracial$ormultiracial$).ti,ab.

8expPrisoners/orprison*.ti,ab.

9exprefugees/or“EmigrantsandImmigrants”/or“TransientsandMigrants”/

10(immigrant*orrefugee*ormigrant*orasylumseeker*).ti,ab.

11expgypsies/ortravel?er*.ti,ab.

12expHomelessYouth/orexpHomelessPersons/orhomeless$.ti,ab.

13expSpouseAbuse/orDomesticViolence/orexpbatteredwomen/

14((abuse$orviolen$)adj4(partner$orwifeorwivesorspouse$ordomestic)).ti,ab.

15((neighbo?rhoodoreconomicorruralorurban)adj2(depriv$orpoverty)).ti,ab.

16(disadvantag*ordeprivedarea*orinnercit*orinnercit*).ti,ab.

17MentalDisorders/orexpeatingdisorders/orexpmooddisorders/orexp“schizophreniaanddisorderswithpsychoticfeatures”/

18((mental$orpsych$)adj2(ill$ordisorder$orimpair$ordisturb$ordisabil$)).ti,ab.

19LearningDisorders/orMentalDeficiency/

20((mental$orlearningorcognitiv$)adj2(retard$orhandicap$ordisab$ordifficult$orimpair$)).ti,ab.

21expProstitution/orsexworker*.ti,ab.

22AdolescentHealthServices/orexpAdolescent/orexpPregnancyinAdolescence/

23(teen$oryouth$oradolescen$).ti,ab.

24expHIVInfections/orHIV/

25(HIVorHIV-pos$orHIV-inf$).ti,ab.

26expStreetDrugs/orexpNarcotics/orexpCocaine/orexpCrackCocaine/orexpHeroin/orexpamphetamines/orexpmethadone/27expsubstance-relateddisorders/orexpSubstanceAbuse,Intravenous/orexpamphetamine-relateddisorders/orexpcocaine-relateddisorders/orexpmarijuanaabuse/orexpopioid-relateddisorders/orexpheroindependence/orexpphencyclidineabuse/orexppsychoses,substance-induced/orexpsubstanceabuse,intravenous/orsubstancewithdrawalsyndrome/

28expalcohol-relateddisorders/orexpalcoholism/orexpalcohol-induceddisorders/

29expCircumcision,Female/


30(femaleadj(genitalmutilationorcircumcisionorgenitalcutting)).ti,ab.

31(clitoridectomyorinfibulation).ti,ab.

32((languageadj3(secondorproblem*oradditionalorbarrier*))ortranslat*orinterpreter*).ti,ab.

33expcommunicationbarriers/orexplanguage/

34expculture/orexpculturalcharacteristics/orexpculturaldiversity/

35((cultur*orsociocultur*orsocio-cultur*)adj5(barrier*ordifferen*orpractice*orsensitiv*orappropriate*)).ti,ab.

36or/1-35

37expPrenatalCare/ormaternalhealthservices/

38((antenatalorprenatal)adj2(careorclinicorprogram*orservice*)).ti,ab.

39expMidwifery/

40or/37-39

41outreach.ti,ab.

42 41 and 40

43((utilis$orutiliz$orbarrier$oraccess$oruptakeorinitiateorinitiationorbooking)adj5(prenatalorantenatalorcare)).ti,ab.

44((lateorearly)adj5(uptakeorinitiat$orattend$orbooking)).ti,ab.

45 (43 or 44) and 40

46PrenatalCare/ut[Utilization]

47 (42 or 45 or 46) and 36

48limit47toinprocess

49 limit 47 to in data review

50limit47topubmednotmedline

51or/48-50

52 47 not 51

53 limit 52 to humans

54 51 or 53 (1622)

55 limit 54 to (english language and yr=”1990 - 2009”)

56casereports/

57(letterorrevieworcommentoreditorialorletterornews).pt.

58 55 not (56 or 57)


Annex B: Named intervention searchesInoursecondroundofsearches,weusedthefollowinglistof(potentiallyeligible)namedinterventions for text searching:

AfterCare Project 69

AsianMotherandBabyCampaign42

Baby Talk 70

California Black Infant Health Program 53

Center for Addiction and Pregnancy71

Community Health Nursing Prenatal Care Program 72

Congress Alukura 62

Daruk Antenatal Program 73

De Madres a Madres74 75 76 77

Florida Outreach Childbirth Education Project 78

HealthStart 52

Homeless Prenatal Program 79

ImprovedPregnancyOutcome80

KansasHealthyStartHomeVisitingProgram48

Maternal Infant Health Advocate Service 50

Maternal Infant Health Outreach Worker 81

Maternal Outreach Worker 82

Minority Health Coalitions’ Early Pregnancy Project 46

MumsandBabiesprogram83

NguaGundi(Mother/ChildProject)61

Omaha Healthy Start 47

OpeningDoors84

Parenting and the Community Health 85

PeerSupportProgramme86

Prenatal Care Assistance Program 56

Project MotherCare87

Resource Mothers Program 55 45

Rural Alabama Pregnancy and Infant Health 88

RuralMaternalChildHealthprogram89

Rural Oregon Minority Prenatal Progam 43

Southeast Asian Health Project 90

Strong Women Strong Babies Strong Culture 41

TeenParentingPartnership91

TempleInfantandParentSupportServices92

The Door 93

Un Comienzo Sano (A Healthy Beginning) 94


Annex C: Characteristics and results of included studiesNotes – how to read this table

Interventiongroupsaredescribedincolumn7.Inmoststudiesthereisonlyone•interventiongroup,labelledI1,denotingInterventiongroup1;wherethereismorethanoneinterventiongroup,groupsarelabelledI1,I2,etc.

Comparator/controlgroup(s)aredescribedincolumn8.Wherethereisonlyone•comparator/controlgroupthisislabelledC1,denotingcontrol/comparatorgroup1;wheretherearemultiplecomparatorgroupsthesearelabelledC1,C2,etc.

Resultsaregenerallypresentedasacomparisonoftheoutcomesintheintervention•groupcomparedwiththecontrolgroup(s),i.e.I1vs.C1forstudieswithoneinterventiongroupandonecontrol/comparatorgroup.Wheretherearemultiplecontrol/comparatorgroups,multiplecomparisonsareshown.

Subgroupanalysesarepresentedwheretheauthorreportsondifferential•effectivenessacrosssubgroups

Bothunadjustedandadjustedresultsarepresentedwhereavailable;wherethe•authorshavefittedmultipleadjustmentmodelswepresenttheresultsconsideredmostrelevant–usuallyinvolvingadjustmentformaternalcharacteristics/riskfactorspresentatbooking.

95%confidenceinterval,‘p-values’and/orastatementthatadifferenceis“not•significant”(NS)areincludedwherereportedbytheauthors.

Forstudieswhichcompareoutcomesbeforeandaftertheimplementationofan•intervention,resultsarepresentedasC1(“before”)vs.I1(“after”)

Abbreviations

ANC=Antenatalcare; OR=Oddsratio; RR=RelativeRisk; 95%CI=95%confidenceinterval; NS=Notstatisticallysignificantatthe5%level.


Au

tho

r an

d

year

Co

un

try/

S

ett

ing

Stu

dy d

esi

gn

Stu

dy p

op

ula

tio

nTarg

et

po

pu

lati

on

(i

nte

rven

tio

n)

Meth

od

of

all

oca

tio

n t

o

stu

dy g

rou

p(s

)In

terv

en

tio

n g

rou

p(s

)C

on

tro

l/

com

para

tor

gro

up

(s)

Resu

lts

– t

imin

g o

f in

itia

tio

n o

f A

NC

Cra

mer

2007

USA,Douglas

County,

Neb

rask

a.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Wom

en g

ivin

g b

irth

in

the

study

area

s bet

wee

n 2

002 a

nd

2003.

Eth

nic

min

ority

wom

en

(“Bla

ck w

om

en”)

. N.B.studysample

not

rest

rict

ed b

y et

hnic

ity)

.

Uncl

ear.

Retrospective

assi

gnm

ent.

I1:

236 w

om

en w

ho

lived

in t

he

19 t

arget

ed

censu

s tr

acts

and w

ho

participatedinthe

programme.

C1:

1520 w

om

en w

ho

lived

in t

he

19 t

arget

ed

censu

s tr

acts

who d

id

nottakepartinthe

programme.

C2:15,949women

who g

ave

birth

in t

he

county

(in

cludes

sm

all

numberofprogramme

participants).

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

2002:I167.1%,C168.0%,C285.5%

2003:I179.0%,C174.5%,C283.3%

I12002vs.C22002,p=0.01

Authorspresentarangeofother

comparisons.SeeTable3inthepaper

for

furt

her

det

ails

.

Daa

lem

an 1

997USA,Topeka-

ShawneeCounty,

Kansas.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Multiparous

wom

en a

ged

17-3

8 r

ecei

ving

WIC

ser

vice

s duringpregnancy,

whohadnoprior

participationinthe

KansasMaternal

and I

nfa

nt

Progra

m

(aseparate

comprehensive

preventative

serv

ice

that

se

rves

hig

h r

isk

child

bea

ring w

om

en

and infa

nts

).

Studyperiodnot

specified.

Soci

oec

onom

ical

ly

dis

adva

nta

ged

wom

en

(“atrisk”,andin

receiptofWIC).

Retrospective

assi

gnm

ent

bas

ed

on c

are

rece

ived

.

I1:?*womenwhohad

participatedinthe

KansasHealthyStart

HomeVisiting(HSHV)

programme(had

rece

ived

at

leas

t one

contact(phonecalland/

orhomevisit))priorto

theindexpregnancy.

*62womenwere

enro

lled in s

tudy

overall,butnumbers

in inte

rven

tion a

nd

comparatorgroupnot

separatelyreported.

Inte

rven

tion t

arget

ed

“at-

risk

” fa

mili

es.

C1:?*womenwhohad

notparticipatedinthe

HSHVprogrammeprior

totheindexpregnancy.

*seenoteinprevious

colu

mn

Numericresultsnotreported;data

derivedvisuallyfrombarcharts,

approximate.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1 v

s. C

1:

63%

vs.

75%

Test

for

tren

d a

cross

thre

e tr

imes

ters

andnocaregroup:p=0.36.

Stratifiedanalysis

Sin

gle

wom

en:

I1 v

s. C

1:

73%

vs.

57%

Test

for

tren

d a

cross

thre

e tr

imes

ters

andnocaregroup:NS.

Edger

ley

2007

USA,PaloAlto,

Cal

iforn

ia.

Low

-inco

me

nei

ghbourh

oods.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Wom

en w

hose

an

tenat

al a

nd

del

iver

y ca

re w

as

funded

by

Med

iCal

(M

edic

aid)

and

who d

eliv

ered

a

single

ton infa

nt

at

Sta

nfo

rd U

niv

ersi

ty

Med

ical

Cen

ter

between01/01/00

and04/07/04.

Soci

oec

onom

ical

ly

dis

adva

nta

ged

wom

en

(Medicaidrecipients).

Sel

ection b

y si

te o

f care/birth.

I1:

108 w

om

en w

ho

initia

ted a

nte

nat

al c

are

on t

he

Wom

en’s

Hea

lth

Van.

C1:

127 r

andom

ly

selectedfrom2,121

wom

en w

ho initia

ted

ante

nat

al c

are

in a

loca

l co

mm

unity

clin

ic.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1 v

s. C

1:

79.6

% v

s. 5

9.8

%

Testfortrendacrossthreetrimesters:p

= 0

.002.

Hunte

2004

USA,Flint,

Gen

esse

e County,Michigan.

Urb

an a

rea.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Afr

ican

-Am

eric

an

wom

en liv

ing in 4

selectedzipcodes

who g

ave

birth

duringthegroup-

specificstudy

periods.

Min

ority

eth

nic

wom

en

(“Afr

ican

-Am

eric

an

wom

en”)

.

Uncl

ear.

Retrospective

assi

gnm

ent

bas

ed

on c

are

rece

ived

?

I1:

111 w

om

en w

ho

wer

e M

IHAS c

lients

and

had

giv

en b

irth

at

the

tim

e of th

e birth

rec

ord

s ex

trac

tion in A

ugust

2003,livinginthe4

targetedzipcodes.

C1:350women,a

“uniformprobability

sample”ofallbirths

duringthefirst6months

of 2003 t

o w

om

en w

ho

wer

e not

MII

HAS c

lients

an

d w

ho liv

ed in 3

of th

e 4targetedzipcodes.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1vs.C1:66.7%vs.67.7%,testof

statisticalsignificancenotreported.


Au

tho

r an

d

year

Co

un

try/

S

ett

ing

Stu

dy d

esi

gn

Stu

dy p

op

ula

tio

nTarg

et

po

pu

lati

on

(i

nte

rven

tio

n)

Meth

od

of

all

oca

tio

n t

o

stu

dy g

rou

p(s

)In

terv

en

tio

n g

rou

p(s

)C

on

tro

l/

com

para

tor

gro

up

(s)

Resu

lts

– t

imin

g o

f in

itia

tio

n o

f A

NC

Jewell2000

USA,Indiana.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

M

atch

ed.

Wom

en w

ho g

ave

birth

during t

he

two y

ear

study

period(datesnot

specified).

Eth

nic

min

ority

wom

en

("m

inority

wom

en")

.U

ncl

ear.

Retrospective

assi

gnm

ent

bas

ed

on c

are

rece

ived

?

I1:

95 w

om

en r

ecei

ving

care

coord

inat

ion t

hro

ugh

enro

lmen

t in

one

of th

e th

ree

Min

ority

Hea

lth

Coalitionprojects.

Inte

rven

tion t

arget

ed a

t w

om

en fro

m m

inority

backgrounds,butsample

not

rest

rict

ed b

y ra

ce.

C1:188women,a

stratifiedrandom

sampleofwomenwho

did

not

rece

ive

care

co

ord

inat

ion.

Eac

h

inte

rven

tion b

irth

was

matchedwithapprox.

2 c

ontr

ols

fro

m b

irth

certificaterecords,

matchedon:race,age,

maritalstatus,and

educa

tion a

ttai

nm

ent.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1 v

s. C

1:

73.3

% v

s. 5

3.3

%

Chi sq

uar

e te

st for

tren

d a

cross

thre

e trimestersandnocaregroup:p=

0.0

10.

Julnes1994

USA,Norfolk,

Virginia.City

with h

igh r

ate

of

infa

nt

mort

ality

and a

dole

scen

t pregnancy.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Adole

scen

ts <

19

year

s w

ho g

ave

birth

during t

he

12 m

onth

stu

dy

period(datesnot

specified).

Teen

ager

s.U

ncl

ear.

Retrospective

assi

gnm

ent

bas

ed

on c

are

rece

ived

?

I1:

49 a

dole

scen

ts

who w

ere

clie

nts

of

the

Res

ourc

e M

oth

ers

Progra

m (

RM

P).

Inte

rven

tion t

arget

ed

“hig

h-r

isk”

adole

scen

ts.

C1:

46 a

dole

scen

ts w

ho

didnotparticipateinthe

RMP(instead,theywere

clie

nts

of th

e cl

inic

-basedmulti-disciplinary

programme).

Unad

just

ed %

initia

ting A

NC b

efore

the

4thmonthofpregnancy:

I1vs.C1:53.1%vs.32.6%,p<0.05

Mac

kerr

as 2

001Australia,Darwin

Rura

l an

d

Eas

t Arn

hei

m

regions,Northern

Terr

itory

.

Bef

ore

and

afte

r st

udy.

Aborigin

al w

om

en

who g

ave

birth

inthethreepilot

com

munitie

s.

Indig

enous

wom

en

(Aborigin

al w

om

en).

Sel

ection b

y ye

ar

ofcare/birth

(pre-andpost-

inte

rven

tion).

I1:

228 w

om

en w

ho

gav

e birth

bet

wee

n

1994and95,afterthe

implementationofthe

Str

ong W

om

en S

trong

Bab

ies

Str

ong C

ulture

(S

WSBSC)

inte

rven

tion.

C1:

246 w

om

en w

ho

gav

e birth

bet

wee

n 1

990

and1991,beforethe

implementationofthe

SW

SBSC inte

rven

tion.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1vs.C1:24.4%vs.16.7%,p=0.03

Mar

tin 1

997

USA,Cincinnati,

Ohio

.Bef

ore

and

afte

r st

udy.

Adole

scen

ts a

ged

14-1

7 y

ears

who

rece

ived

ante

nat

al

carethroughGroup

Hea

lth A

ssoci

ates

.

Teen

ager

s.Sel

ection b

y ye

ar

ofcare/birth

(pre-andpost-

inte

rven

tion).

I1:

72 a

dole

scen

ts w

ho

gav

e birth

bet

wee

n 1

992

and1994,receivingtheir

ante

nat

al c

are

afte

r th

e implementationofthe

Teen

Pre

gnan

cy P

rogra

m

(TPP

).

C1:

33 a

dole

scen

ts

whogavebirth1991,

rece

ivin

g in t

hei

r an

tenat

al c

are

bef

ore

theimplementation

oftheTPP.Allprivate

patients(clinicdidnot

acceptMedicaiduntil

1992).

Unad

just

ed %

initia

ting A

NC b

efore

4th

monthofpregnancy:

I1vs.C1:79.8%vs.69.7%,NS.

Mas

on 1

990

U.K,Leicester.

Two m

ater

nity

units

and f

our

GP

surg

erie

s.

Prospective

obse

rvat

ional

co

hort

stu

dy.

Asi

an w

om

en

regis

tering for

ante

nat

al c

are

at

selectedpractices

between01/05/85

and30/04/86.

Eth

nic

min

ority

wom

en

(Gujarati,Punjabi,

Hindi,Urduand

Bengalispeakers).

Uncl

ear. N

on-

random

sel

ection.

I1:

213 w

om

en w

ho h

ad

at lea

st o

ne

conta

ct w

ith

acommunityand/or

hospitallinkworker.

C1:

244 w

om

en w

ho d

id

not

hav

e an

y co

nta

ct

withacommunityand/

orhospitallinkworker.

Unad

just

ed %

initia

ting A

NC b

efore

12

wee

ks g

esta

tion:

I1vs.C1:70%vs.70%,NS

Morr

is 1

993

USA,Galveston,

Texa

s. C

linic

s providedbythe

Univ

ersi

ty o

f Te

xas

Med

ical

Bra

nch

.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

Adole

scen

ts a

ged

<

18 y

ears

who

rece

ived

ante

nat

al

care

at

the

study

clin

ics

and g

ave

birth

bet

wee

n

1985 a

nd 1

986.

The

maj

ority

of

adole

scen

ts w

ere

med

ical

ly indig

ent.

Teen

ager

s.Sel

ection b

y si

te

ofcare/birth.Self-

sele

ctio

n.

I1:

660 a

dole

scen

ts w

ho

rece

ived

ante

nat

al c

are

attheteenpregnancy

clin

ic.

C1:

277 a

dole

scen

ts

who r

ecei

ved “

trad

itio

nal

prenatalcare”.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1 v

s. C

1:

45.2

% v

s. 1

9.5

%

Distributionoftrimesterofbooking,chi

squaretest:p=0.001.


Au

tho

r an

d

year

Co

un

try/

S

ett

ing

Stu

dy d

esi

gn

Stu

dy p

op

ula

tio

nTarg

et

po

pu

lati

on

(i

nte

rven

tio

n)

Meth

od

of

all

oca

tio

n t

o

stu

dy g

rou

p(s

)In

terv

en

tio

n g

rou

p(s

)C

on

tro

l/

com

para

tor

gro

up

(s)

Resu

lts

– t

imin

g o

f in

itia

tio

n o

f A

NC

Mvu

la 1

998

USA,New

Orleans,

Louis

iana.

Pu

blic

ly

funded

clin

ics

serv

ing “

low

so

cioec

onom

ic

populations,

with m

ost

bel

ow

povertylevel”.

Prospective

obse

rvat

ional

co

hort

stu

dy.

Med

ical

ly low

-ris

k w

om

en r

egis

tering

for

ante

nat

al c

are

between01/01/94

and31/12/94,

excl

udin

g w

om

en

withmultiple

pregnancies

andspecific

complications

(includingdiabetes,

hypertension,HIV,

sick

le c

ell)

.

Soci

oec

onom

ical

ly

dis

adva

nta

ged

wom

en

(low

inco

me)

.

Sel

ection b

y si

te o

f care/birth.

I1:

179 w

om

en

who r

egis

tere

d for

ante

nat

al c

are

at

the

Nei

ghbourh

ood

Preg

nan

cy C

are

(NPC

).

C1:

181 r

andom

ly

sele

cted

wom

en w

ho

regis

tere

d for

ante

nat

al

care

at

the

Louis

iana

Sta

te U

niv

ersi

ty

obst

etric

clin

ic.

Unad

just

ed %

initia

ting A

NC b

efore

14

wee

ks:

I1vs.C1:37%vs.29%,

Distributionoftrimesterofbooking,chi

squar

e te

st:

NS

New

schaf

fer

1998

USA,NewYork

Sta

te.

Med

icai

d

Ante

nat

al C

linic

s.

Retrospective

obse

rvat

ional

co

hort

stu

dy.

HIV-positive

subst

ance

abusi

ng

wom

en c

laim

ing

Med

icai

d w

ho g

ave

birth

to a

sin

gle

ton

infa

nt

bet

wee

n

January1993and

September1994.

Soci

oec

onom

ical

ly

dis

adva

nta

ged

wom

en

(Medicaidrecipients),

though inte

rven

tion

evaluatedinHIV-

positivesubstance

abusi

ng w

om

en.

Uncl

ear. S

elec

tion

bysiteofcare/

birth

?

I1:

240 w

om

en w

ho

rece

ived

ante

nat

al

care

at

a Pr

enat

al C

are

Ass

ista

nce

Pro

gra

m

(PCAP)

clin

ic (

wom

en

with a

t le

ast

one

Med

icai

d P

CAP

clai

m

duringpregnancy).

C1:

113 w

om

en w

ho

rece

ived

ante

nat

al c

are

from

a s

ourc

e oth

er

than

PCAP.

Unad

just

ed %

initia

ting A

NC b

efore

the


I1vs.C1:76%vs.68%,p=0.11.

Rei

chm

an 1

996

USA,NewJersey.Retrospective

obse

rvat

ional

co

hort

stu

dy.

Med

icai

d c

laim

ants

w

ho h

ad a

sin

gle

ton

live

birth

bet

wee

n

1989 a

nd 1

990.

Soci

oec

onom

ical

ly

dis

adva

nta

ged

wom

en

(Medicaidrecipients).

Sel

ection b

y si

te o

f care/birth

11:24,036women

(10,908Black,13,128

White)whoparticipated

inNewJersey’s

HealthStartprogramme.

C1:16,719women

(8,671Black,8,102

White)

who d

id n

ot

participateinNew

Jersey’sHealthStart

programme.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

Onlystratifieddatareportedbyauthors.

Bla

cks:

I1

vs.

C1:

49.5

% v

s. 4

9.6

% (

NS)

Whites

: I1vs.C1:56.4%vs.59.0%,p<0.01

Roger

s 1996

USA,South

Car

olin

a. R

ura

l (<

50 %

urb

an)

and m

oder

atel

y urb

an (

bet

wee

n

50-7

5%

urb

an)

counties

.

Retrospective

obse

rvat

ional

co

hort

st

udy

with

additio

nal

pre-

inte

rven

tion

comparator

group.

Primiparous

adole

scen

ts (

<18

year

s at

del

iver

y)

who g

ave

birth

to

a s

ingle

infa

nt

between01/01/86

and31/12/89.

Teen

ager

s.U

ncl

ear.

Retrospective

assi

gnm

ent

bas

ed

on c

are

rece

ived

?

I1:

1901 a

dole

scen

ts

whoparticipatedin

the

Res

ourc

e M

oth

ers

Progra

m (

RM

P) a

nd

resi

ded

in t

he

16

counties

in w

hic

h t

he

RMPwasimplemented.

C1:

4612 a

dole

scen

ts

who r

esid

ed in 1

6

comparisoncounties

(bro

adly

mat

ched

to

the

RM

P co

unties

on

selectedsociocultural,

perinatalstatus

and h

ealth r

esourc

e in

dic

ators

) w

ho d

id n

ot

takepartintheRMP

programme.

C2:

712 a

dole

scen

ts

who r

esid

ed in 1

0 o

f th

e 16 R

MP

counties

an

d g

ave

birth

bef

ore

theRMPprogramme

wasimplemented

in t

hei

r co

unty

of

resi

den

ce (

inte

rven

tion

wasimplementedina

stepwisefashion).

Unad

just

ed %

initia

ting A

NC b

efore

the


I145.3%,C140.9%,C240.0%

Adjusted*oddsratioforearlyinitiation

of AN

C (

1-3

month

s)

I1vs.C1,OR1.48(1.32,1.66)

I1vs.C2,OR1.39(1.16,1.66)

*adjustedforage,maritalstatus,race,

andpreviouspregnancies


Au

tho

r an

d

year

Co

un

try/

S

ett

ing

Stu

dy d

esi

gn

Stu

dy p

op

ula

tio

nTarg

et

po

pu

lati

on

(i

nte

rven

tio

n)

Meth

od

of

all

oca

tio

n t

o

stu

dy g

rou

p(s

)In

terv

en

tio

n g

rou

p(s

)C

on

tro

l/

com

para

tor

gro

up

(s)

Resu

lts

– t

imin

g o

f in

itia

tio

n o

f A

NC

Thompson1998USA,Oregon.

“Veryrural

county

”.

Bef

ore

and

afte

r st

udy.

Lo

w-i

nco

me

Mex

ican

- Am

eric

an

wom

en liv

ing in

rura

l ar

eas.

Eth

nic

min

ority

wom

en

(“M

exic

an-A

mer

ican

w

om

en”)

.

Sel

ection b

y ye

ar

ofcare/birth

(pre-andpost-

inte

rven

tion).

I1:100highrisk*

pregnantwomen

recr

uited

bet

wee

n

September1991-May

1994 w

ho r

ecei

ved

Rura

l O

regon M

inority

Pr

enat

al P

rogra

m

(RO

MPP

) se

rvic

es a

nd

wer

e co

nsi

der

ed t

o

havecompletedthe

inte

rven

tion (

rece

ived

a

min

imum

of 3 v

isits)

.

*highrisk:included

historyofpreterm

orLBW,historyof

pregnancycomplications,

pre-existingmedical

conditions,age<17

year

s.

C1:

100 w

om

en

“demographically

sim

ilar”

to inte

rven

tion

samplewhogavebirth

in t

he

study

county

bet

wee

n 1

989-9

1

(the

year

s bef

ore

the

RO

MPP

inte

rven

tion w

as

implemented).

Numericresultsnotreported;

approximatedataderivedvisuallyfrom

bar

char

ts.

Unad

just

ed %

initia

ting A

NC b

efore

the


I1 v

s. C

1:

66%

vs.

52%

Distributionofgroupedmonthof

booking(1-2,3-4,5-6,7+ornocare),

Man

n-W

hitney

tes

t: N

S.

Will

is 2

004

USA,California.

Prospective

obse

rvat

ional

co

hort

stu

dy.

Afr

ican

-Am

eric

an

Med

icai

d c

laim

ants

w

ho g

ave

birth

during r

elev

ant

studyperiodand

resi

ded

in t

arget

ed

ZIPcodes.

Eth

nic

min

ority

wom

en

(“Afr

ican

-Am

eric

an

wom

en”)

.

Uncl

ear.

I1:2,031womenwho

had

a s

ingle

ton liv

e birth

betweenJuly1996and

September1990and

whoparticipatedinthe

Bla

ck I

nfa

nt

Hea

lth (

BIH

) programme.

C1:11,622womenwho

gav

e birth

in 1

997 a

nd

didnotparticipateinthe

BIHprogramme.

Unad

just

ed %

initia

ting A

NC in 1

st

trim

este

r:

I1vs.C1:60.8%vs.75.8%,testof

statisticalsignificancenotreported.

Please cite this document as:

LauraOakley,RonGray,JenniferJKurinczuk,PeterBrocklehurst,JenniferHollowell,Asystematic review of the effectiveness of interventions to increase the early initiation of antenatal care in socially disadvantaged and vulnerable women. Oxford: National Perinatal EpidemiologyUnit,2009

ThisreportandotherInequalitiesinInfantMortalityProjectreportsareavailableat:

www.npeu.ox.ac.uk/infant-mortality

a systematic review of the effectiveness of interventions to

Documents