Use of a Prenatal Risk Screen to Predict Maternal Traumatic Pregnancy-Associated Death: Program and Policy Implications

MISSING IMAGE

Material Information

Title:
Use of a Prenatal Risk Screen to Predict Maternal Traumatic Pregnancy-Associated Death: Program and Policy Implications
Series Title:
Hardt N, J Eliazar, M Burt, R Das, WP Winter, H Saliba, and J Roth. “Use of a prenatal risk screen to predict maternal traumatic pregnancy-associated death: program and policy implications” Women’s Health Issues May-June 2013; 23(3): e187-e193 http://dx.doi.org/10.1016/j.whi.2013.02.002
Physical Description:
Journal Article
Creator:
Hardt, Nancy S
Eliazar, Jessica
Burt, Martha J
Das, Rajeeb
Winter, William P
Saliba, Heidi
Roth, Jeffrey
Publisher:
Elsevier (Women's Health Issues)
Publication Date:

Notes

Abstract:
Background: Motor vehicle crashes, homicide, suicide, and drug abuse are among the leading causes of pregnancy- associated deaths. To prevent such deaths, identifying women for intervention is required. The universally offered Florida Healthy Start Prenatal Risk Screen was evaluated to identify women at increased risk for traumatic pregnancy- associated death. Methods: Florida ’ s Enhanced Maternal Mortality Reporting Database for 1999 through 2005 was linked with Florida ’ s Healthy Start Prenatal Risk Screen to identify traumatic pregnancy-associated death as the outcome. Distribution of Healthy Start risk scores among women who died were compared with the screened population. Traumatic death esti- matesper100,000birthsweredrawnforeachriskscore,alongwithestimatesoftherelativerisk(RR)oftraumaticdeathfor each score. The RR of womenwith scores greater than or equal to 4 were compared with the risk of women scoring 0 to 3. Findings: Almost 20% of the 620,959 women who did not die of traumatic death had a risk score of 0, compared with only 3% of the 144 women who did die of traumatic death. As risk scores increased, the chance of traumatic deaths sharply increased. A woman with a score of 4 had 11.78 times (con fi dence interval CI, 4.63 – 29.69) the risk of traumatic death compared with a woman with a risk score of 0. Conclusions: The implementation of prenatal risk screening to identify women at increased risk for traumatic pregnancy- associated death would help to ensure that policies to reduce infant risk factors also address maternal risk factors
Acquisition:
Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Haeden Roberson.
Publication Status:
Published

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the submitter.
System ID:
IR00004273:00001


This item is only available as the following downloads:


Full Text

PAGE 1

OriginalarticleUseofaPrenatalRiskScreentoPredictMaternalTraumatic Pregnancy-AssociatedDeath:ProgramandPolicyImplicationsNancyS.Hardt,MDa,b,JessicaEliazar,MPHc,MarthaBurt,MDa,RajeebDas,MSPHd, WilliamP.Winter,MPHe,HeidiSaliba,BAf,JeffreyRoth,PhDd,g,*aDepartmentofPathology,UniversityofFlorida,Gainesville,FloridabDepartmentofObstetrics/Gynecology,UniversityofFlorida,Gainesville,FloridacTruvenHealthAnalytics,AnnArbor,MichigandOf ceofInstitutionalPlanningandResearch,UniversityofFlorida,Gainesville,FloridaeEighthJudicialCircuitofFloridaDomesticViolenceFatalityReviewTeam,Of ceoftheMedicalExaminer,District8,Gainesville,FloridafPed-I-CareandtheDivisionofGeneralPediatrics,UniversityofFlorida,Gainesville,FloridagDepartmentofPediatrics,UniversityofFlorida,Gainesville,Florida Articlehistory: Received24August2012;Receivedinrevisedform15February2013;Accepted27February2013abstractBackground: Motorvehiclecrashes,homicide,suicide,anddrugabuseareamongtheleadingcausesofpregnancyassociateddeaths.Topreventsuchdeaths,identifyingwomenforinterventionisrequired.Theuniversallyoffered FloridaHealthyStartPrenatalRiskScreenwasevaluatedtoidentifywomenatincreasedriskfortraumaticpregnancyassociateddeath. Methods: Florida sEnhancedMaternalMortalityReportingDatabasefor1999through2005waslinkedwithFlorida s HealthyStartPrenatalRiskScreentoidentifytraumaticpregnancy-associateddeathastheoutcome.Distributionof HealthyStartriskscoresamongwomenwhodiedwerecomparedwiththescreenedpopulation.Traumaticdeathestimatesper100,000birthsweredrawnforeachriskscore,alongwithestimatesoftherelativerisk(RR)oftraumaticdeathfor eachscore.TheRRofwomenwithscoresgreaterthanorequalto4werecomparedwiththeriskofwomenscoring0to3. Findings: Almost20%ofthe620,959womenwhodidnotdieoftraumaticdeathhadariskscoreof0,comparedwith only3%ofthe144womenwhodiddieoftraumaticdeath.Asriskscoresincreased,thechanceoftraumaticdeaths sharplyincreased.Awomanwithascoreof4had11.78times(con denceinterval[CI],4.63 29.69)theriskoftraumatic deathcomparedwithawomanwithariskscoreof0. Conclusions: Theimplementationofprenatalriskscreeningtoidentifywomenatincreasedriskfortraumaticpregnancyassociateddeathwouldhelptoensurethatpoliciestoreduceinfantriskfactorsalsoaddressmaternalriskfactors. Copyright 2013bytheJacobsInstituteofWomen sHealth.PublishedbyElsevierInc.Introduction Overthelastcentury,maternalmortalityhassigni cantly decreasedintheUnitedStatesandinmanyothercountries ( Changetal.,2003 ).Althoughthereductioninratesofmaternal deathsfrompregnancy-relatedcausesseemstohaveleveledoff ( Hoyert,2007 ),maternaldeathsfromtraumaticcausesnow accountforabout46%ofallmaternaldeaths( Mirza,Devine,& Gaddipati,2010 ).Traumaticcausesofdeathincludemotor vehiclecrashes,homicide,suicide,accidentalpoisoning,and othertypesofinjuries.Therateoftraumaticdeathsamong pregnantandrecentlypregnantwomen,combinedwiththe historicaldeclineinpregnancy-relatedcausesofdeath,hasledto ashiftintheproportionofdeathswithinthispopulationfrom pregnancy-relatedtotraumaticcauses( Romero&Pearlman, 2012 ). Traumaticcausesofdeathareoftenexcludedfrommaternal deathreviewsbecausetheyarenotconsideredtobedirectly relatedtothepregnancy.Forthissamereason,theyarealsonot SupportedinpartbyagrantfromMarthaBurt,DepartmentofPathology, UniversityofFloridatoJessicaEliazarforaninternshipattheFamilyDataCenter, DepartmentofPediatrics,UniversityofFlorida. Correspondenceto:JeffreyRoth,PhD,UniversityofFlorida,Departmentof Pediatrics,POBox100296,Gainesville,FL32610-0296,UnitedStates.Phone: 1 3523341360;fax: 13523341361. E-mailaddress: rothj@peds.u .edu (J.Roth). www.whijournal.com1049-3867/$-seefrontmatterCopyright 2013bytheJacobsInstituteofWomen sHealth.PublishedbyElsevierInc. http://dx.doi.org/10.1016/j.whi.2013.02.002Women'sHealthIssues23-3(2013)e187 e193

PAGE 2

generallyincludedinstandardde nitionsandmeasurementsof maternalmortality.Astheproportionoftraumaticcausesof maternaldeathincreases,exclusionofthesewomenfrom maternalmortalityreviewhindersaccurateestimateofrisks;it alsodelayseffectiveinterventions( Horon,2005 ; Horon&Cheng, 2011 ). Traumahasnowemergedasaleadingcauseofpregnancyassociateddeath.Onestudyofdeathsamongpregnantand postpartumwomeninCookCounty,Illinois,foundthat46% diedfromtraumaticcauses,comparedwith32%from pregnancy-relatedcauses( Fildes,Reed,Jones,Martin,&Barrett, 1992 ).AsimilarstudyinNewYorkCityattributed39%of pregnancy-associateddeathstotrauma( Dannenbergetal., 1995 ).AstatewidestudyinGeorgiaontraumaticpregnancyassociateddeathhadcomparableresults, ndingthat35%of pregnancy-associateddeathswereattributabletotraumatic causes( Dietz,Rochat,Thompson,Berg,&Grif n,1998 ). Figure1 displaysthetop10causesofpregnancy-associateddeathin Florida,rankedbypregnancy-associatedmaternalmortality ratio,fortheyears1999through2005.Thetopfourcausesare traumaticinnature( Hernandez,Sappen eld,&Burch,2009 ).In Florida,motorvehicleaccidents,homicide,accidental poisoning,andsuicidearetheleadingcausesoftraumatic pregnancy-associateddeath. Kavanaughandcolleaguesrecentlydocumentedtherelationshipbetweenpsychosocialriskfactors(suchassubstance abuse,mentalillness,anddomesticviolence)andtraumatic pregnancy-associateddeathsinVirginia( Kavanaughetal., 2009).Theyfoundthatthepregnancy-associatedmaternal mortalityratioforwomenexperiencingoneormorepsychosocialriskfactorswas17.1deathsper100,000livebirths, comparedwiththeoverallpregnancy-relatedmortalityratioof 11.8deathsper100,000reportedfortheU.S.population between1991and1999( Changetal.,2003 ).Substanceabuse contributedto29%ofallcases,mentalillnessto17%,and domesticviolenceto14%.Unintentionalinjuries,which includedunintentionaloverdosesandmotorvehiclecrashes, werethemostcommoncauseoftraumaticpregnancyassociateddeathamongwomenwithoneormorerisk factors,followedbyhomicideandsuicide. The AmericanCollegeofObstetriciansandGynecologists (2006) recommendspsychosocialscreeningforallwomen seekingprenatalcare,regardlessofsocialstatus,educational level,orrace/ethnicity.Multiplestudiesontraumaticpregnancyassociateddeathalsorecommendtheuseofsomeformof psychosocialscreeningasanessentialcomponentofcomprehensivemedicalcareforwomenbothduringandafter pregnancy(e.g.,Krulewitch,Pierre-Louis,deLeon-Gomez,Guy,& Green,2001 ; Martin,Beaumont&Kupper,2003 ; Parsons& Harper,1999 ; Shadigian&Bauer,2005 ).Manyofthepsychosocialriskfactorsassociatedwithadversefetalandbirthoutcomes, suchassubstanceabuse,domesticviolence,andmentalillness, arealsoassociatedwithanincreasedriskoftraumatic pregnancy-associateddeath( Behrman&Butler,2007 ).InFlorida,thereisnostandardizedscreeningtoolforidentifying womenatanincreasedriskoftraumaticpregnancy-associated death.However,Floridastatuterequiresthatallprenatalcare providersofferariskscreentowomenatthe rstprenatalcare visitandagainatthetimeofchildbirth( FloridaStatute383.14 ). Thisscreenincludesdemographicaswellasenvironmentaland socialfactorstoidentifywomenatriskforpretermdelivery ( Table1 ). Florida sHealthyStartprogramwascreatedin1991bythe statelegislaturewiththegoalofreducingpretermbirthand otherhigh-riskconditions.Theprogramprovidesservicesto pregnantwomenandmotherswithnewbornsthatinclude mentalhealthcare,substanceabusecounseling,carecoordination,infantsafetyeducation,breastfeedingsupport,smoking cessation,andinterconceptionalcare( Clark,Watson,Thompson, &Sappen eld,2006 ).Thestatuterequiresobstetricprovidersto completetheriskassessmentofdemographic,environmental, andpsychosocialfactorssothatwomencanbereferredto appropriatehealth,education,andsocialservices.Manyofthe riskfactorsidenti edthroughtheuseofthisscreeningtoolare alsoriskfactorsassociatedwithtraumaticpregnancy-associated death.Ariskscoreof4orgreaterisusedbyprenatalcare providerstoidentifywomenandinfantseligibleforHealthyStart services.Servicesincludepsychosocial,nutritional,andsmoking cessationcounseling.Inadditiontotheseservices,childbirth education,breastfeedingeducation,andparentingeducationare deliveredthroughhomevisitation. TheFloridaDepartmentofHealthcontractswith33Healthy StartCoalitionstoadministertheseservices.Thecoalitions comprisebothpublicandprivateproviders.Florida sstatewide, decentralizedapproachisdistinctfromthefederalHealthyStart initiativewhichalsostartedin1991andtargetsspeci c communitieswithlargedisparitiesininfantbirthoutcomes ( Taylor&Nies,2012 ).ThreecountiesinFlorida(Hillsborough, Duval,andPinellas)havebeenrecipientsoffederalHealthy StartfundingfromtheMaternalandChildHealthBureau.An evaluationofthefederalCentralHillsboroughCountyHealthy StartprograminTampa,Florida,foundthatprogramparticipantshadloweroddsoflowbirthweightandpretermbirth comparedwithnon-participants( Salihu,Mbah,Jeffers,Alio,& Berry,2009 ).Thisevaluationexaminedtheimpactoflocal pre-andpost-natalriskreductionservicestowomeninfour EastTampazipcodeswhohadcompletedtheHealthyStart prenatalriskscreen.Thepresentstudydoesnotdistinguish womenwhoreceivedHealthyStartservicesthroughtheir participationinoneofFlorida sthreefederalHealthyStart initiatives;norweredataavailableaboutthetypeofservices receivedbyHealthyStart-screenedwomen. Becausecertainconditionsthatplaceawomanatriskfor adverseinfantbirthoutcomesarelikelytoalsoputheratriskof traumaticdeath,wehypothesizedthattheprenatalriskscreen mightalsobeusefultoidentifyriskoftraumaticpregnancyassociateddeath.Thepurposeofthisstudywastoevaluatethe potentialuseoftheFloridaHealthyStartPrenatalRiskScreenas atooltoidentifywomenatincreasedriskfortraumatic pregnancy-associateddeath. Figure1. Top10causesofpregnancy-associateddeathinFlorida,excludingnatural causesrankedbymaternalmortalityratio. N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e188

PAGE 3

Methods Casesofpregnancy-associateddeathwereidenti edbythe FloridaDepartmentofHealth senhancedmaternalmortality reportsurveillancesystem.Fouridenti cationcriteriawereused inanattempttomaximizetheidenti cationofpregnancyassociateddeathswithinthestate:1)Thedeathcerti cate responsetothequestion, Wasthewomanpregnantwithinthe pastyear? wasmarked, yes ;or2)the InternationalClassi cationofDiseases ,TenthRevision(ICD-10)codeindicatedadeath classi edasbeingowingtocausesrelatedto, Pregnancy, Childbirth,andthePuerperium (O00 O99);or3)therewas amatchingfetalbirthordeathrecordon leforthewoman within365daysbeforeherdeath;or4)therewasamatching HealthyStartPrenatalScreenon leforthewomanwithin365 daysbeforeherdeath.Thismethodofpregnancy-associated deathidenti cationisamongthemostcomprehensiveinthe UnitedStates( Burch,Noell,Hill,&Delke,2012 ; King,2012 ). Thenextstepintheenhancedmaternalmortalitysurveillancesystemistoreviewthematernaldeathcerti catesofall identi edpregnancy-associateddeathsandassignthemtoone ofthreecategoriesbasedontheprimarycauseandmannerof death.The rstcategory, pregnancy-relateddeath ,isde nedas, apregnancy-associateddeathresultingfrom1)complications ofthepregnancyitself,2)thechainofeventsinitiatedbythe pregnancythatledtodeath,or3)aggravationofanunrelated conditionbythephysiologicorpharmacologiceffectsofthe pregnancythatsubsequentlycauseddeath ( Watson,Thompson, Burch,&Sappen eld,2008 ,p.3).Thesecondcategory, possible pregnancy-relateddeath ,isde nedas, apregnancy-associated deathwheredeterminationofthedeathcouldnotbeconclusivelyclassi edaseitherrelatedornotrelatedtothepregnancy (Ibid).Thethirdcategory, notpregnancyrelated ,isde nedas, the deathofawoman,whilepregnantorwithinoneyearof terminationofpregnancy,fromacausedeemedunrelatedto pregnancy (Ibid). Traumaticpregnancy-associateddeathsarenottypically abstractedforreviewbythemultidisciplinarymaternal mortalitysurveillanceteam.TheFloridaDepartmentofHealth providedtheFamilyDataCenterattheUniversityofFlorida witha lecontaininginformationontraumaticpregnancyassociateddeathsthatoccurredbetweentheyearsof1999 and2005. DataontheHealthyStartPrenatalRiskScreenwasrequested fromtheFloridaDepartmentofHealthforthisstudy.Duringthe studyperiod,therewere1,542,055births.Ofthose,404,788 womendidnothavearecordedscreen,leaving1,137,267women screened.Reasonswomenwerenotscreenedincludedlackof prenatalcare,refusingtheofferedscreeningtest,andprovider notofferingthescreenattheprenatalvisitoratthetimeofbirth. Incasesofsuccessivedeliveriesbythesamewomanduringthe studyperiod,onlythemostrecentscreenwasretained,leaving 674,875screens.Thisnonduplicationofcaseswasintendedto capturethescreensperformedclosesttothetimeofmaternal death,therebymaintainingaconsistentmethodologyofone observationperdeath.Recordswithmissingvaluesonany HealthyStartriskscreenquestionweredeleted( n 53,772), leaving621,103completerecordsforanalysis. Table2 compares theproportionsofsociodemographiccharacteristicsamong womenwhowereandwerenotscreened. Thetraumaticpregnancy-associateddeathdatasetwasthen linkedwiththeHealthyStartPrenatalRiskScreendatasetto identifywomenwhohaddiedoftraumaticcauseswithin1year oftheterminationoftheirpregnancy.Theunmatchedrecords andrecordswithinvalidsocialsecuritynumbers( n 142,747) werelinked,allowingfor60%ofpreviouslyunmatchedrecords tobelinkedbasedonmultiplematchingentriesoncommon demographic eldssuchasdateofbirth,race,andzipcode.A traumaticdeathindicatorwasthencreatedwithintheHealthy StartPrenatalRiskScreendatasettoallowfortheidenti cation oftraumaticdeathasanoutcome. Table3 comparesthe proportionsofsociodemographiccharacteristicsamongHealthy Table1 FloridaHealthyStartPrenatalRiskScreeningFactorsandRiskScoringPoints Answers Points Questionsansweredbypatient Yourage (inyears) < 18 1pointor > 39 1point Yourrace Black/White/other Black 2points*Areyoumarried? Yes/No No 1point HaveyougraduatedfromhighschoolorreceivedaGED? Yes/No No 1point Yourweightbeforepregnancy. (inpounds) < 110 1point Doyouhaveanyproblemswhichpreventyoufromkeepingyourhealthcareorsocial servicesappointments? Yes/No Yes 1point Haveyoumovedmorethan3timesinthelast12months? Yes/No Yes 1point Doyoufeelunsafewhereyoulive? Yes/No Yes 1point Doyouoranymemberofyourhouseholdgotobedhungry? Yes/No Yes 1point Inthelast2months,haveyouusedanyformoftobacco? Yes/No Yes 1point Inthelast2months,haveyouuseddrugsoralcohol(includingbeer,wine,mixeddrinks)?(a)earlier,(b)later, (c)notatall,(d)nochange Yes 1point Ifyoucouldchangethetimingofthispregnancy,wouldyouwantit (a)earlier,(b)later, (c)notatall,or(d)nochange (c)Notatall 1point Questionsansweredbyhealthcareprovider Didpatient slastpregnancyresultinamiscarriage,stillbirth,ababylessthan51/2 pounds,ababybornmorethan3weeksearly,orababythatstayedinthehospitalafter thepatientwenthome? Yes/No Yes 1point Doespatienthaveanyillnessthatrequirescontinuingmedicalcare? Yes(specifyillness)/No Yes 1point Trimesterofentryat rstprenatalvisit First/second/third Secondtrimester 1point Pointsaretotaledforeachrespondent.Atotalscoreof4ormoreisconsideredapositivescreening.*Blackraceisscored2pointsbecausetheriskofadversebirthoutcome(lowbirthweight < 2000gandpretermdelivery < 34weeks)isdoublethatobservedin womenofanyotherraceinFlorida.InApril2007,FloridareviseditsHealthyStartPrenatalRiskscreen.Acopyofthecurrentscreenisavailableat: http://www.doh.state. .us/family/mch/hs/english_prenatal_screen.pdf .N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e189

PAGE 4

Start-screenedwomenwhodiedtraumaticallyandunscreened womenwhodiedtraumatically. TheHealthyStartPrenatalRiskScreenscorewascalculated foreachwomanbysummingthepointsrecordedforthe15risk factors.Blackraceisscored2pointsbecausetheriskofadverse birthoutcome(lowbirthweight[ < 2,000g]andpreterm delivery[ < 34weeks])isdoublethatobservedinwomenofany otherraceinFlorida( Simmons,Thompson,&Graham,2003 ). WithBlackracereceiving2points,thehighestpossiblescoreis 16.Nosubjectshadscoresmorethan13andnotraumaticdeaths wereobservedamongwomenwhoscored10ormore. Comparisonofgroupswasthereforerestrictedtosubjectswith scoreslessthanorequalto9.Thenumberoftraumatic pregnancy-associateddeathsper100,000atriskwasestimated foreachofthe10riskgroupscorrespondingtoscores0to9. Relativerisks(RR)alongwiththeircon denceintervals(CI)were estimatedforcomparingrisksbetweenthefollowingsubgroups: 1)Pregnantwomenineachofthenineriskgroupswithrisk scoresbetween1and9werecomparedwiththosewithrisk score0;2)pregnantwomenineachofthenineriskgroupswere comparedwiththoseinthenextlowerriskgroup;and3) pregnantwomenwithriskscoresof4orhigherwerecompared withthosewithriskscoresof0to3.ThoseRRswerecompared withthecurrentuseofthisscreentoidentifyriskofadversefetal andinfantoutcomes.CIswerecalculatedatthe95%con dence levelusingthedeltamethodforestimatingthestandarderrors ( Agresti,1990 ).ThisstudywasapprovedbytheInstitutional ReviewBoardsoftheFloridaDepartmentofHealthandthe UniversityofFlorida. Results Between1999and2005,therewere620,959womenwho completedtheHealthyStartPrenatalRiskScreenand144 experiencedapregnancy-associatedtraumaticdeath. Figure2 showscontrastingdistributionsofscreenscoresforwomen whosufferedtraumaticdeathversuswomenwhodidnot. Almost20%ofthe620,959womenwhodidnotdieoftraumaticdeathhadariskscoreof0,comparedwithonly3%ofthe 144womenwhodiddieoftraumaticdeath.Fifty-sixpercent ofthewomenwhodiedoftraumahadariskscoregreaterthan orequalto4,comparedwith28%ofwomenwhodidnotdieof trauma. Figure3 showsanincreaseinpredictedtraumatic deathsper100,000birthsandacorrespondingincreaseinRR asHealthyStartRiskScoresincrease. Table4 showsthe numberofpredictedpregnancy-associatedtraumaticdeaths per100,000birthsforriskscoresbetween0and9.For example,among100,000pregnantwomenwithariskscoreof 4,wecanexpect47.88traumaticdeaths(CI,45.59 50.28).The graphin Figure3 showsadramaticincreaseinpredicted traumaticdeathsper100,000birthsasHealthyStartRisk Scoresincrease. Becauseariskscoreof4isanimportantcutoffforreferralto services,weshowin Table5 thatawomanwithascoreof4has 11.78times(CI,4.63 29.69)theriskoftraumaticdeath comparedwithawomanwithariskscoreof0. Table6 shows thattheestimatedRRandits95%CIforapregnantwomanwith agivenscore(s)comparedwithawomanwiththenextlower score( s 1).Withascoreof4,awomanhasapredictedriskof traumaticdeath1.79timesthatofawomanwithscoreof3(CI, 1.11 2.89). TheRRofpregnancy-associatedtraumaticdeathforapregnantwomanwithriskscoreof4ormorewascalculatedtobe 3.27(CI,2.35 4.56)comparedwithwomenwhoscored0to3. ThisRRisnearlydoublethatreportedforaHealthyRiskPrenatal Table3 ComparisonofWomenWhoDiedTraumaticallyinFlorida,1999 2005:Screened andNotScreenedbyHealthyStart Demographic Characteristic Percentageof WomenWhoDied Traumatically andScreenedby HealthyStart Percentageof WomenWhoDied Traumatically andNotScreenedby HealthyStart Maternalage(yrs)*< 2033.019.7 20 34 53.9 51.1 > 35 13.2 29.3 Maritalstatus*Married 23.0 55.3 Notmarried77.044.7 Maternaleducation*Highschool graduate d yes 53.462.8 Highschool graduate dno 46.6 37.2 RecordsreceivedfromtheFloridaDepartmentofHealth senhancedmaternal mortalityreportsurveillancesystemdidnotcontaininformationaboutdecedents race.*Indicateschi-squaretestdeterminedthattheproportionsinthelevelsofthe demographiccharacteristicwerestatisticallydifferentinthetwogroups. Table2 ComparisonofWomenWhoDeliveredinFlorida,1999 2005:ScreenedandNot ScreenedbyHealthyStart DemographicCharacteristicPercentageof WomenScreened byHealthyStart PercentageofWomen NotScreenedby HealthyStart Maternalage*< 2014.27.7 20 34 74.7 72.8 > 35 11.1 19.4 Maternalrace*White 69.4 77.0 Black 25.4 17.9 Other 5.2 5.1 Maritalstatus*Married 53.5 70.7 Notmarried 46.5 29.3 Maternaleducation*Highschoolgraduate dyes39.5 53.0 Highschoolgraduate dno60.5 47.0*Indicateschi-squaretestdeterminedthattheproportionsinthelevelsofthe demographiccharacteristicwerestatisticallydifferentinthetwogroups. Figure2. Distributionofriskscoresonthehealthystartprenatalscreen. N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e190

PAGE 5

RiskScreenof4orgreatertopredictadversebirthoutcome (RR,1.77; Clark&Thompson,2004 ). Inadditiontoitsef cacyinidenti cationofriskforinfant mortality( FloridaDepartmentofHealth,2010 )andlowbirth weight( Thompson,2011 ),theFloridaHealthyStartscreen seemstobeusefulfortheidenti cationofwomenatincreased riskoftraumaticdeath.Itislikelythat,becausethisscreen includessimilarquestionstoprenatalriskscreensusedin otherstates,the ndingsfromthisstudymaybegeneralizedto otherproviderswhosescreensassessthe15riskfactorsnamed herein. Discussion Thestrengthofthisstudyisthatitusesalarge,administrative datasettoanalyzeaninstrumentthatisrequiredbystatutetobe usedtoscreeneverypregnantwoman.Alimitationofthestudy isthatitisretrospective,andbene tstowomenofinterventions resultingfromthescreeningarespeculative.Itiscurrently unknowntowhatextentinterventionsintendedtoreduce prematurityalsohavebene cialeffectsonmothers risksof traumaticdeath.Indeed,becausetheservicesprovidedto eligiblemothersvarydependingoncapacityoflocalservice providers,itisalimitationofthisstudytoassociateanyspeci c serviceprovidedtowomenwiththeiroutcomeintermsof maternaldeath. Pregnancy-associatedtraumaticdeathisarareevent.In Florida,theincidenceis24traumaticdeathsperyearonaverage. Althoughnotallwomenparticipate,Medicaidbene ciariesare morelikelytocompletetheHealthyStartScreen( Florida DepartmentofHealth,2007 ).However,thehighestriskgroup ofall d womenwhodeliverwithoutreceivinganyprenatal care d were,byde nition,unscreened. Acomparisonoftheproportionsinthesociodemographic characteristicsofwomenwhowerescreenedandnotscreened indicatedthatthetwogroupsdifferedintheirsociodemographic characteristics( Table2 ).Womenwhowerenotscreenedwere lesslikelytobeteenagers,morelikelytobemarried,andmore likelytobehighschoolgraduates.Then,acomparisonofthe proportionsinthesociodemographiccharacteristicsof unscreenedandscreenedwomenwhodiedoftraumarevealed thesamedifferences.Unscreenedwomensufferingtraumatic deathwerelesslikelytobeteenagers,morelikelytobemarried, andmorelikelytobehighschoolgraduates. AlthoughallwomeninFloridaareofferedtheHealthyStart Screen,providersaremorelikelytoobtainpatientpermission toscreenwhentheyperceivethatwomenwouldqualifyfor andbene tfromHealthyStartservices.These ndingsillustratetheimportanceofuniversalscreeningforrisk.Providers cannottellbylookingattheirpatientswhomightbene tfrom screeningandservices.Indeed,itisinterestingtospeculate whethersomeoftheunscreenedwomen sdeathsmighthave beenpreventedifreferralforHealthyStartserviceshadbeen made. ImplicationsforPracticeandPolicy Amajorityofpregnantandpostpartumwomenareseen repeatedlybydoctorsorotherhealthcareprofessionalsduring thecourseoftheirpregnancyandinthepostpartumperiod. Thesevisitscreateauniqueopportunityforthedetectionofrisk factorsandsubsequent,appropriateinterventionstoreduce them.Intimatepartnerviolence,prescribedorrecreationaldrug useormisuse,motorvehiclesafetyincludinguseofseatbelts andavoidingtextinganddrivingarepotentialissuesthatcould beaddressedduringprenatalandpostpartumcare.Theimplementation ofprenatalriskscreentoalsoidentifywomenat increasedriskfortraumaticpregnancy-associateddeathwould helptoensurethatappropriatereferralsaremadeforthe protectionofbothmotherandbaby. WomeninFloridawhoscore4ormoreontheHealthyStart Prenatalscreenareofferedhomevisitationservicesbyanurse. Acceptanceofofferedservicesisnotrequired,butthosewho doacceptreceiveaseriesofvisitswherehealthandsocial needsareaddressed.Althoughitisunknownwhetherservices offeredinhopesofimprovinginfantoutcomesactually improvedmaternaloutcomes,itisplausiblethatinvolvement inprenatalhealthpromotionactivitieslowerstheriskof traumaticdeath.Forexample,womenwhodonotfeelsafeat homearefurtherscreenedandreferredforservicesiftheyare victimsofintimatepartnerviolence.Intimatepartnerviolence threatensthewell-beingofinfantandmotherequally.Women needingsafehousingmaybeassistedinrelocatingtohousing forwhichtheyqualify.Saferhousingforinfantsislikelyto bene tmothersaswell.Womenwithinsuf cientfoodtomeet theirfamilies nutritionalneedsreceiveassistanceand encouragementtobreastfeed.Womenwithmedicalcomplicationsofpregnancyreceivesupportensuringaccessto follow-upvisitsandrecommendedtherapies.Womenwho smokeareassistedwithaccesstosmokingcessationsupport. Mostimportant,at-riskwomendeveloparelationshipwith thehomevisitationprovider. ThestrengthoftheHealthyStartprogramresidesinthe trustingrelationshipbuiltbetweenanonjudgmentalprovider andclientorpatient.Whendeveloped,astrongandlongFigure3. Predictedtraumaticdeathsper100,000birthsandcorrespondingrelative risk. Table4 PredictedNumberofTraumaticDeathsper100,000Pregnanciesand95%Con denceInterval RiskScore 0 123456789 No.ofdeaths 4.099.1718.2726.7647.8830.3549.0771.84126.14215.05 Uppercon dencelimit6.0511.1520.1528.6350.2834.5857.7795.03205.78571.79 Lowercon dencelimit2.767.5416.5625.0145.5926.6341.6854.3177.3380.88N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e191

PAGE 6

lastingrelationshipoffersopportunitiestoemphasizepublic healthmeasuresthatmayreducepregnancy-associated maternalmortalityowingtotrauma.Historically,theserelationshipswerecreatedbetweenthepatient/clientandthe providerduringahomevisit,butequallyeffectiverelationshipsmaybecreatedatothersites,suchaswork,faith-based institutions,orcommunitycenters.BecauseHealthyStart servicescontinueafterthebabyisborn,theprovidermay modelparentingbehaviortothebene tofthemotherand otherfamilymembers.Theprovidercanalsobevigilantin assessingforpostpartumdepression(whichmayleadto accidentaldrugoverdoseorsuicide).Riskreductionservices couldbeexpandedtoincludeeducationalinterventionsthat maylimitthepossibilityoftraumaticinjurytomothers.For example,providerscanremindmothertofastenherseatbelt, useachildcarseat,andavoidcellphoneortextmessaging whiledriving.Theprovidermayalsoprovidecounselingonthe dangersofdrinkinganddriving,orgettingintoacarwithan impaireddriver.Thesesafetyreminderscouldalsobecome partofof cecaredeliveredduringpregnancyandpostpartum visits. Physiciansandotherwomen shealthprofessionalsare avitallinktoresourcessuchasHealthyStart.Iftheresultsof theHealthyStartscreenbecameanadditionalvitalsign recordedatthe rstprenatalvisit,addressingthesocial determinantsofhealthwouldbecomeanessentialpartof women shealthcare( Halfon,Larson,&Russ,2010 ).Obstetric visitscanandoftendoextendbeyondmedicalcaretoaddress underlyingpsychosocialissuesandrisks.Inthecaseofpregnantwomen,sometraumaticpregnancy-associateddeaths maybepreventableifprovidersareawareofthedoubleutility oftheFloridaHealthStartPrenatalRiskScreen.Programsand policiesthatcurrentlyfocusoninfantoutcomescould,inthe future,addressmorebroadlythelifecourseoutcomesof womenandchildren.BecausetheHealthyStartPrenatalRisk Screenhaselementsthatidentifybothmotherandinfantat riskforprematuremortality,theoverlapinriskfactorsis noteworthy.Indeed,asecondinstrumenttoassessmaternal riskoftraumaticdeathislikelynotrequired. BecausetheAffordableCareActhascreatedasetof preventionandscreeningmeasuresthatwillbeavailabletoall womeninthefutureandwillbecoveredbyinsuranceplans, morewomenwillundergointimatepartnerviolencescreening attheirhealthcarevisits.Thisscreeningcouldincludeother socialriskfactorsascoveredbytheHealthyStartScreen withoutaddingundulytothelengthofthescreening encounter,redoublingthevalueofthescreeningandallowing forappropriatereferralandintervention. Acknowledgments TheauthorsthankstaffmembersoftheFloridaDepartment ofHealth,BureauofCommunityHealthAssessmentfor approvingtheuseofandmakingavailablebirthvitalrecordsand HealthyStartscreensforthisresearch:DeborahBurch,Karen Freeman,MeadeGrigg,JoAnneSteele;andstaffoftheOf ceof ReproductiveHealthwhocommentedonanearlyversionofthis paper:SheronikaDenson,WilliamSappen eld,ErinSauberSchatz,andDanThompson. TheauthorsalsothankcolleaguesattheUniversityofFlorida whometregularlytoreviewearlyversionsofthispaper:ErikW. Black,OnyekachukwuOsakwe,PVRao,andLindsayThompson. EspeciallyhelpfulwereinsightsprovidedbyKarenE.Harris, PresidentofNorthFloridaWomen sPhysiciansofGainesville, LLC. ReferencesAgresti,A.(1990). Categoricaldataanalysis .NewYork:Wiley-Interscience. AmericanCollegeofObstetriciansandGynecologists.(2006).ACOGCommittee OpinionNo.343:Psychosocialriskfactors:Perinatalscreeningandintervention. ObstetricsandGynecology ,108(2),469 477. Behrman,R.E.,&Butler,A.S.(Eds.).(2007). Pretermbirth:Causes,consequences, andprevention ..Washington,DC:NationalAcademiesPress. Burch,D.,Noell,D.,Hill,W.C.,&Delke,I.(2012).Pregnancy-associatedmortality review:TheFloridaexperience. SeminarsinPerinatology ,36(1),31 36. Chang,J.,Elam-Evans,L.D.,Berg,C.J.,Herndon,J.,Flowers,J.,Seed,K.A.,etal. (2003).Pregnancy-relatedmortalitysurveillance dUnitedStates,1991 1999. MorbidityandMortalityWeeklyReport ,52(2),1 8. Clark,C.,&Thompson,D.(2004).FactorsassociatedwithofferratesforFlorida s HealthyStartprenatalscreen:Ananalysisofthe2001deliverycohort. RetrievedAugust20,2012,from http://www.doh.state. .us/Family/mch/ docs/pdf/PrenatalScreen.pdf Clark,C.,Watson,A.,Thompson,D.,&Sappen eld,W.(2006).MaternalcharacteristicsofwomenwhoarescreenedandreferredtotheFloridaHealthy StartPrenatalProgrambasedonfactorsotherthanprenatalscreenrisk score.RetrievedAugust20,2012,from http://www.doh.state. .us/Family/ mch/docs/BOOF_White_Paper.pdf Dannenberg,A.L.,Carter,D.M.,Lawson,H.W.,Ashton,D.M.,Dorfman,S.F.,& Graham,E.H.(1995).Homicideandotherinjuriesascausesofmaternal deathinNewYorkCity,1987through1991. AmericanJournalofObstetrics andGynecology ,172(5),1557 1564. Dietz,P.M.,Rochat,R.W.,Thompson,B.L.,Berg,C.J.,&Grif n,G.W.(1998). Differencesintheriskofhomicideandotherfatalinjuriesbetweenpostpartumwomenandotherwomenofchildbearingage:Implicationsfor prevention. AmericanJournalofPublicHealth ,88(4),641 643. Fildes,J.,Reed,L.,Jones,N.,Martin,M.,&Barrett,J.(1992).Trauma:Theleading causeofmaternaldeath. JournalofTrauma,32(5),643 645. FloridaDepartmentofHealth.(2007).FloridaHealthyStartPrenatalRiskScreen. RetrievedAugust20,2012,from http://www.doh.state. .us/family/mch/hs/ english_prenatal_screen.pdf FloridaDepartmentofHealth.(2010).HealthyStartfactsheet.Retrieved August20,2012,from http://www.doh.state. .us/family/mch/hs/hs_ factsheet_2_10.pdf FloridaStatute383.14.Screeningformetabolicdisorders,otherhereditaryand congenitaldisorders,andenvironmentalriskfactors.RetrievedAugust20, 2012,from http://www.lawserver.com/law/state/ orida/statutes/ orida_ statutes_383-14 Table5 RelativeRisksand95%Con denceIntervalsforSubjectsinSpeci edRiskCategoriesComparedwithZero HealthyStartScore 1 2 3 4 5 6 7 8 9 Relativerisk 2.24 4.47 6.5511.78 7.4312.0117.58 30.87 52.63 Uppercon dencelimit6.5711.9116.9229.6920.4434.0955.39114.92270.96 Lowercon dencelimit0.77 1.68 2.54 4.63 2.7 4.23 5.58 8.29 10.22 Table6 EstimatedRelativeRisks(RR)and95%Con denceIntervalsforComparing SubjectsinSpeci cRiskCategories RiskScore 123 456789 RR(s:s-1) 2.241.991.461.790.631.621.461.761.79 Lowercon dencelimit0.770.930.831.110.350.760.580.510.31 Uppercon dencelimit6.574.262.592.891.153.453.725.999.28N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e192

PAGE 7

Halfon,N.,Larson,K.,&Russ,S.(2010).Whysocialdeterminants? Healthcare Quarterly ,14(1),8 20. Hernandez,L.,Sappen eld,W.M.,&Burch,D.(2009).Florida spregnancyassociatedmortalityreview2009update.RetrievedAugust20,2012,from http://www.doh.state. .us/Family/mch/pamr/2009_PAMR_Update.pdf Horon,I.L.(2005).Underreportingofmaternaldeathsondeathcerti catesand themagnitudeoftheproblemofmaternalmortality. AmericanJournalof PublicHealth ,95(3),478 482. Horon,I.L.,&Cheng,D.(2011).Effectivenessofpregnancycheckboxesondeath certi catesinidentifyingpregnancy-associatedmortality. PublicHealth Reports ,126(2),195 200. Hoyert,D.L.(2007).Maternalmortalityandrelatedconcepts. VitalHealth Statistics ,3(33),1 13. Kavanaugh,V.M.,Fierro,M.F.,Suttle,D.E.,Heyl,P.S.,Bendheim,S.H.,& Powell,V.(2009).PsychosocialriskfactorsascontributorstopregnancyassociateddeathinVirginia,1999 2001. JournalofWomensHealth(Larchmont) ,18(7),1041 1048. King,J.C.(2012).MaternalmortalityintheUnitedStates dWhyisitimportant andwhatarewedoingaboutit? SeminarsinPerinatology ,36(1),14 18. Krulewitch,C.J.,Pierre-Louis,M.L.,deLeon-Gomez,R.,Guy,R.,&Green,R. (2001).Hiddenfromview:Violentdeathsamongpregnantwomeninthe DistrictofColumbia,1988 1996. JournalofMidwiferyandWomensHealth 46(1),4 10. Martin,S.L.,Beaumont,J.L.,&Kupper,L.L.(2003).Substanceusebeforeand duringpregnancy:Linkstointimatepartnerviolence. AmericanJournalof DrugAlcoholAbuse ,29(3),599 617. Mirza,F.G.,Devine,P.C.,&Gaddipati,S.(2010).Traumainpregnancy:A systematicapproach. AmericanJournalofPerinatology ,27(7),579 586. Parsons,L.H.,&Harper,M.A.(1999).ViolentmaternaldeathsinNorthCarolina. ObstetricsandGynecology ,94(6),990 993. Romero,V.C.,&Pearlman,M.(2012).Maternalmortalityduetotrauma. SeminarsinPerinatology ,36(1),60 67. Salihu,H.M.,Mbah,A.K.,Jeffers,D.,Alio,A.P.,&Berry,L.(2009).HealthyStart programandfeto-infantmorbidityoutcomes:Evaluationofprogrameffectiveness. MaternalandChildHealthJournal ,13(1),56 65. Shadigian,E.,&Bauer,S.T.(2005).Pregnancy-associateddeath:Aqualitative systematicreviewofhomicideandsuicide. ObstetricalandGynecological Survey ,60(3),183 190. Simmons,M.,Thompson,D.,&Graham,C.(2003).AnevaluationoftheHealthy StartPrenatalScreen1998birthcohort.RetrievedAugust20,2012,from http://www.doh.state. .us/Family/mch/docs/prenatal_screen.pdf Taylor,Y.J.,&Nies,M.A.(2012).Measuringtheimpactofoutcomesofmaternal childhealthfederalprograms. MaternalandChildHealthJournal .Jun23. [Epubaheadofprint]. Thompson,D.(2011).HealthyStartPrenatalScreening:Pretermbirthandlowbirth weightpercentagesbyscreeningscore.RetrievedAugust20,2012,from http:// www.doh.state. .us/Family/mch/docs/HealthyStartPrenatalScreeningrates0927-11revised04-02-12.pdf Watson,A.,Thompson,D.,Burch,D.,&Sappen eld,B.(2008). Pregnancy-related mortalityreport,Florida1999 2005 .Tallahassee:FloridaDepartmentof Health.AuthorDescriptionsNancyS.Hardt,MD,isProfessorofPath ologyandObstetrics-Gynecologyat theUniversityofFloridaCollegeofMedicine,andisamemberofFlorida s Pregnancy-AssociatedMortalityReviewteam. JessicaEliazar,MPH,iscurrentlyServiceDeliverymanagerforTruvenHealth Analytics,AnnArbor,Michigan. MarthaBurt,MD,isAssistantProfessorofPathologyattheUniversityofFlorida CollegeofMedicineandDistrict8MedicalExaminer. RajeebDas,MSPH,isSeniorProgramEvaluator,Of ceofInstitutionalPlanningand Research,UniversityofFlorida,Gainesville,Florida. WilliamP.Winter,MPH,isCoordinatorofEighthJudicialCircuitofFloridaDomestic ViolenceFatalityReviewTeam. HeidiSaliba,BA,isResearchCoordinatorforPed-I-CareandtheDivisionofGeneral Pediatrics,UniversityofFlorida. JeffreyRoth,PhD,isResearchProfessorofPediatricsattheUniversityofFloridaand ProgramDirectoroftheFamilyDataCenter. N.S.Hardtetal./Women'sHealthIssues23-3(2013)e187 e193 e193