Accident Analysis and Prevention

ContentslistsavailableatScienceDirect

Accident

Analysis

and

Prevention

jo u r n al hom e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / a a p

Road

traffic

injury

incidence

and

crash

characteristics

in

Dar

es

Salaam:

A

population

based

study

Karen

Zimmerman

_,

_Ali

_A.

_Mzige

_,

_Pascience

_L.

_Kibatala

_,

_Lawrence

_M.

_Museru

_,

_Alejandro

_Guerrero

a_{Amend,P.O.Box152,NewYork,NY10101,USA}

b_{InternationalMedicalandTechnologicalUniversity,DaresSalaam,Tanzania} c_{MinistryofHealthandSocialWelfare,DaresSalaam,Tanzania}

d_{MuhimbiliOrthopaedicInstitute,DaresSalaam,Tanzania} e_{IntertraumaMedicalConsulting,NewYork,NY,USA}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received31March2011

Receivedinrevisedform23June2011 Accepted28June2011

Keywords: Roadtrafficinjuries Injuryprevention Tanzania

Traumaandpopulation-basedinjury incidence

a

b

s

t

r

a

c

t

Roadtrafficinjuries(RTI)areapublichealththreatandamajorsourceofdisabilityindevelopingcountries. Apopulation-basedanalysisofRTIsinatestimoniallyhigh-riskareaofDaresSalaam,thelargestcity intheEastAfricancountryofTanzania,wascarriedoutwiththegoalofestablishinganRTIincidence andtoidentifyRTIcharacteristicsthatmaybeusedforatargetedinjurypreventionprograminthese communities.

Geographicclustersamplingwascompletedin2adjacentwardsofDaresSalaamwithhousehold sur-veysadministeredinpersontodetermineadenominator.AnyhouseholdmembersinvolvedinanRTI withintheprevious12monthsreceivedanin-depthquestionnaire.Demographics,incident character-istics,medicalattention,injuriesanddisabilitydayswerenoted.Thesearedescribedandcomparedto injuryseverityandagespecifictendencies.

Withinthe30clusters,6001individualswereinterviewed.Ofthem,196wereinvolvedinnon-fatal RTIswithintheprevious12months,resultinginanon-fatalincidencerateof32.7RTIsper1000person years.Therewere4deathsnoted.Injuriesresultinginafracturecorrelatedwithadisabilityofmorethan 30days.Childrenwereinjuredaspedestrians93%ofthetimeandweremorelikelytobeinjuredonsmall, unpavedsidestreetsthanadults.MostRTIsoccurredonahighwayandaffectedthelowerextremities, requiredtreatmentatahospital,andresultedinapolicereportbeingfiled50.2%ofthetime.

Inconclusion,RTIsinthisurbanEastAfricansettingareamajorsourceofdisability.Thisstudyprovides incidencedataandcrashcharacteristicsthatmaybeusedtoconstructpreventionprogramsandcould validatesecondarydatasources.

© 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Roadtrafficinjuries(RTI)areamajorpublichealththreatand withoutpreventativemeasuresareprojectedtoincrease signifi-cantlyworldwideoverthenext20years(Pedenetal.,2004).RTIs accountforthelargestproportionofunintentionalinjuriesandare increasinglyrecognizedinlow-incomecountriesasamajorcause ofmorbidityandmortality(Chandranetal.,2010;Pedenetal.,2004, 2009).TheWorldHealthOrganization(WHO)projectsthatRTIs worldwidewillbeoneoftheleadingcausesofdisabilityadjusted lifeyears(DALYs)in2030(Pedenetal.,2004).Despite discourag-ingstatisticssuchasthese,anincreaseinattentionandresearch maybeabletoaltertheincreasingrateofRTIs.Forexample,the safecommunities’ modelhasdemonstratedan injuryreduction

∗Correspondingauthor.Tel.:+12152629819.

E-mailaddress:[email protected](K.Zimmerman).

insomestudies(Spinksetal.,2005).Certainly,speedbumpsand infrastructuredevelopmenthavealsodemonstratedpromise,as haveseatbeltandmotorcyclehelmetlegislation(Dinh-Zarretal.,

2001;MacphersonandMacarthur,2002;Redelmeieretal.,2003;

Servadeietal.,2003;Shultsetal.,2001).However,thereareserious

obstaclesthatquestionthefeasibilityofthesestrategiesin devel-opingcountries.Alowercostoption,suchaschildhoodeducation, mayproveasuccessfulapproachinsuchcountries.

A tool for gathering research,such as creating a successful injury surveillance system,is needed to provide accurate data forpublichealth interventionsand preventionmeasures.These surveillancesystemsprovidethenumbersand typesof injuries that occur as well as the circumstances of the injuries. While thisinformationisreadilyavailableinmostdevelopedcountries intheformofregionallyaggregatedtraumaregistries,its imple-mentationrequires significantinvestment.Thishighcostleaves low-incomecountries’surveillancesystemsextremelylimitedor non-existent.Inaddition,thesecondarydatathatareavailablein 0001-4575/$–seefrontmatter© 2011 Elsevier Ltd. All rights reserved.

Pleasecitethisarticleinpressas:Zimmerman,K.,etal.,RoadtrafficinjuryincidenceandcrashcharacteristicsinDaresSalaam:Apopulation

Keymessages

Whatisalreadyknownaboutthesubject:

•Roadtrafficinjuriesareamajorsourceofmorbidityin devel-opingcountries.

•Secondarydatasourcesareofinconsistentvalue. •ChildrenareparticularlyvulnerabletoRTIs. Whatthisstudyadds:

•ApopulationbasedRTI incidence,which canbe usedfor comparisontoothercommunities.

•Specificcrashand injurycharacteristics valuable for con-structingatargetedpublichealthintervention.

•Childrenandadultshavedifferentcrashcharacteristics. •50%ofRTIvictimsfiledapolicereport.

low-incomecountriesrarelyhavepopulationbaseddatatovalidate

them(Dandonaetal.,2008).

Apopulation-basedstudyoninjuriesbyMoshiroetal.(2001)

foundthat between1992 and 1998transport related accidents weretheleadingcauseofinjuryinDaresSalaam,Tanzania.Despite RTIsbeingshownastheleadingcauseofinjuriesinDaresSalaam, fewstudieshaveillustratedspecificRTIincidenceandcrash char-acteristics(Moshiroetal.,2005).SinceRTIsarealeadingcauseof injury,itisimportanttoquantifytheRTIincidenceandunderstand specificcrashcharacteristics.

Amend isa non-governmental organizationwiththegoalof decreasingroadtraffic injuryrates thoughadvocacy,education, social marketing, and scientific research in Africa. Amend was responsibleforfundingthisstudy,thoughalltheauthors partic-ipatedonavoluntarybasis.Thisstudywasconceivedanddesigned toprovideobjectiveinformationforthedevelopmentofaninjury preventionstrategyinthistestimoniallyhigh-riskareaofDares Salaam.UnderstandingDaresSalaam’sRTI impactand identify-ing specificcrashcharacteristicsis important inrecognizing its subsequentimpactonthecommunityandmayprovidevaluable informationforconstructingpreventionmeasures.

2. Methods 2.1. Studysetting

ThestudytookplaceintheAzimioandMtoniwardsofDares Salaam,thelargestcityandcommercialcapitalofTanzania.Thetwo wardsareadjacentandhaveasinglecommonhighwaybisecting themandwerethereforetreatedasasinglegeographicarea.This areawaschosenbecauseoftestimoniallyhighRTIrates.

2.2. Samplingstrategy

A single-stagecluster samplingwasusedtoselect individu-als and households for an interview. Because density statistics werenotavailable,thestudywascarriedoutwithoutregardfor populationdensity.Thissamplingstrategyisusedwidelyin low-incomecountrieswhereaccuratedataonspecificaddresslocations isnotavailable(Hendersonetal.,1973;HendersonandSundaresan,

1982;Kobusingyeetal.,2001;Moshiroetal.,2001,2005).Inthe

twowards,atotalof30globalpositioningsatellite(GPS)points wererandomlyselected.Byapplyingagridtoasatellitemapofthe studyareaandusingarandomnumbergenerator,coordinateson thegridwereselectedandconvertedtoformalcoordinatesusing

GoogleMapssoftware(GoogleTM_{,MountainView,California).Each} GPSpointwastermedacluster.

Datawascollectedon200individualsclosesttotheactualGPS coordinatesateachcluster.Thesamplesizeof6000individuals forthetwowardscombinedwasdesiredbecauseofthefollowing assumptions:iftheincidencewastakentobe30per1000person years,andwesoughta50%reduction,theninordertoachieve sig-nificancewithadesigneffectof2.0,wewouldrequireasamplesize ofn=2670for80%powerand95%confidenceina2-tailanalysis. Furthermore,adesigneffectcoefficientof2.0wasused,sincethere wasnomannertoquantifyorestimatetheintraclustervariability. Thisvaluehasbeencitedbyotherauthorsasareasonableestimate intheabsenceofapre-existingderivation(Bennettetal.,2002;

HayesandBennett,1999;Hendersonetal.,1973;Hendersonand

Sundaresan,1982).

2.3. Interviewprocess

Research assistants were hired from the local allied health schooltoperformtheinterviews.Thestudentswereselectedafter undergoinganexaminationonthestudyprotocol.Theresearch assistantsweretakentoaspecificclustereachdaytocollectthe data.Ifanyoftheinterviewedindividualsreportedbeinginvolved inanRTIintheprevious12months,a2-pagequestionnairewas administeredintherelevantlanguage.Anindividualwas consid-eredinvolvedinanRTIiftheintervieweestatedthattheindividual hadbeeninvolvedinanRTI.Therewasnodiscriminationfor num-berofdisabilitydays.

Thequestionnairesoughttogatherthefollowinginformation; demographics,circumstancesoftheincident,healthconsequences, long-termfunctionalstatus,economicimpactandlengthof disabil-ity.Informationwasalsocollectedonanyhouseholdmembersthat mayhavedied.Ahouseholdmemberwasconsideredany individ-ualspendingthemajorityofnightsatalocationwithaprimary entrancesharedbytheotherhouseholdmembersoverthe previ-ous12months.Theprincipalinvestigatorreviewedeachcompleted questionnairewiththeresearchassistants,andrandomlyaudited 10%oftheclusterstoensureaccuracy.

The studywaspilottestedtoensuretheinterview methods andquestionnairewasreasonableandproblemfree.Minor adjust-mentsweremadetoensuretheaccuracyandconsistencyofthe interviewprocess.TheprojectwasapprovedbytheTanzanian Min-istry ofHealth andSocial Welfare and theNationalInstituteof MedicalResearch.

2.4. Datamanagement

ThedatawasenteredintoaStatisticsProgramfortheSocial Sci-encesversion18.0database(SPSSinc,Chicago,IL)bytheresearch team.Demographicswerecalculatedforthedenominator,andan injuryincidencewastabulated.Frequenciesandmeanswere calcu-latedforcategoricalandcontinuousvariables,respectively.Minor injuriesweredefinedasthosewithdisabilitydayslessthanorequal to30days,andmajorinjuries,asthosewithgreaterthan30 dis-ability days.IfanindividualwasstillrecoveringfromanRTI at thetimeofinterview,thetimefromtheRTIwasusedasdisability days.Forthepurposesofanalysis,childrenundertheageof1were considered1-yearold.

InordertoappreciatetheeconomicimpactofRTIsinthis com-munity,disabilitydayswereaveragedand summedintotaland for eachagegroupwithoutregardforseverity.Whilerecallhas beenfoundtobevariable,andmostaccuratefor minorinjuries within3months,allinjuriesfromtheprevious12monthswere includedtogetthebroadestdescriptionofinjurycharacteristics andcircumstances(Mocketal.,1999a,b).

Fig.1.Adultandchildcrashcircumstances.

Inferentialanalysiswasperformedtoidentifycrash character-istics,whichcorrelatedwithincreaseddisabilitydays.Fisherexact testwasusedforcontingencyanalysisincategoricalvariablesand ANOVAusedforcomparingmultiplegroups.Significancewastaken tobep<0.05.

Inadditiontoinjuryincidence,theadjustedincidenceratewas calculatedforanyindividualsthatlostatleast1dayofnormal activ-ity.Thiswasdonetoallowcomparisontootherstudies,whichused thisdefinition(Mocketal.,1999b;Moshiroetal.,2001).

Ofnote,therewere4fatalRTIs.Thesewereexcludedfromthe analysisanddescription,becausethedataavailablewithregards totheseinjurieswasdeterminedtobecontradictoryand incon-sistent.Additionallythelownmadeestimationofamortalityrate inaccurate.

3. Results

Datafrom6001individualsintheAzimioandMtoniwardswas collectedforthedenominatorandmadeupthetotalsamplesize. Therewere196individuals(CI:169.9–222.1)whoreporteda non-fatalRTIintheprevious12months.

TheRTIswereseparatedbyagebracketstoreflectpreschool (0–4),schoolaged(5–14),workingage(15–44),andolder(>45). TheagebracketspecificinjuryincidencesaredescribedinTable1. Thetotalincidenceofnon-fatalRTIwas32.7per1000personyears. Inthegreaterthan45agebracketthisfigurewas48.0per1000 personyears.

Fig.1demonstratesthecircumstancesrelatedtoeachRTI, sepa-ratedbychildren(ages0–14)andadults(ages>14).Bicycle-related

Table1

Injuryincidenceanddisabilityaveragesbyagegroup.

Ages 0–4 5–14 15–44 >45 Total RTI within last 12 months 13 29 122 32 196 (6.0–20.0) (18.6–30.4) (100.6–143.4) (21.2–42.8) (169.1–222.1) Denominatora _{780 1314 3241 666 6001} Malesex 38.5% 51.7% 68.0% 59.4% 62.2% Incidenceb _{16.7 22.1 37.6 48.0 32.7} (7.0–25.7) (14.2–30.0) (31.0–44.2) (31.8–64.2) (28.2–37.2) Adjustedb,c _{5.1 16.0 28.0 45.0 24.2} (0–12.1) (9.2–22.8) (20.8–35.2) (29.3–60.7) (20.3–28.1) Disability daysd,e 34.40±93.79 25.19±58.97 49.21±85.58 73.28±85.57 49.19±77.56 n=10 n=26 n=107 n=32 n=175

a_{Denominatorrepresentstotalnumbersurveyedforeachagegroup.} b_{Expressedin1000personyears.}

c_{Onlyincludesthosewithinthelastyearthatmissedatleast1dayofnormalactivity.} d_{Disabilityaveragesincludethosewithatleast1disabilityday.}

Pleasecitethisarticleinpressas:Zimmerman,K.,etal.,RoadtrafficinjuryincidenceandcrashcharacteristicsinDaresSalaam:Apopulation

Table2

Crashcharacteristicscomparedtoinjuryseverity.

Minora _{Major Total p}b

Crashtype

Hitbymini-bustaxi 18 5 23 0.62

9.2% 2.6% 11.7%

Hitbycar 20 15 35 0.11

10.2% 7.7% 17.9%

Hitbytaxi 2 1 3 n/a

1.0% 0.5% 1.5%

Hitbymotorcyle 14 6 20 n/a

7.1% 3.1% 10.2%

Hitbybicycle 22 4 26 0.16

11.2% 2.0% 13.3%

Injuredinmini-bustaxi 21 9 30 n/a

10.7% 4.6% 15.3%

Injuredwhileinacar 15 9 24 0.48

7.7% 4.6% 12.2%

Injuredwhileridinga taxi 3 1 4 n/a 1.5% .5% 2.0% Injuredwhileona motorcycle 15 5 20 0.80 7.7% 2.6% 10.2%

Injuredwhileridinga bicycle 7 1 8 0.44 3.6% 0.5% 4.1% Other 2 2 n/a 1.0% 1.0% Hitby3-wheeled vehicle 1 1 n/a 0.5% 0.5% Circumst. Playing 16 16 19 0.02 8.2% 1.5% 9.7% Walkingtoschool 5 1 6 0.68 2.6% 0.5% 3.1%

Walkingfromschool 7 2 9 n/a

3.6% 1.0% 4.6%

Walkingtowork 8 1 9 0.45

4.1% 0.5% 4.6%

Walkingfromwork 14 5 19 n/a

7.1% 2.6% 9.7%

Walkingelsewhere 22 14 36 0.68

11.2% 7.1% 18.4%

Ridingtoschool 1 1 n/a

0.5% 0.5%

Ridingfromschool 3 3 n/a

1.5% 1.5%

Ridingtowork 18 7 25 n/a

9.2% 3.6% 12.8%

Ridingfromwork 10 6 16 0.57

5.1% 3.1% 8.2% Riding-other 26 10 36 n/a 13.3% 5.1% 18.4% Workingasadriver 5 4 9 0.46 2.6% 2.0% 4.6% Workingasaseller 4 4 8 0.24 2.0% 2.0% 4.1% Total 139 57 196 70.9% 29.1% 100.0%

a_{Majorinjuriesarethosewithgreaterthan30disabilitydays.} b_{n/arepresentseitheranextremelylown,orp=1.0.}

injuries werefoundtobethesamein boththechildand adult groups,comprising13.3%ofallRTIs.Mini-bustaxiswere impli-catedinthecaseofadults30.5%ofthetime,thoughonly14.3%of thetimeinchildren(47/154vs6/42,p=0.11).Privatevehicleswere implicatedin30.1%ofinjuries, withsimilartrendsinboth chil-drenandadults(10/42vs49/154,p=0.58).Theseverityofinjury, asdefinedbydisabilitydays,wascomparedtoRTIcircumstances andanatomicinjuryinTable2forallindividuals.Majorinjuries, definedasadisabilityofgreaterthan30days,represented29.1%of allinjuries.Individualswhomissedatleast1dayofnormalactivity represented74.0%,whereas5.6%expectedtoneverbeabletoreturn toworkorschoolasaresultoftheRTI.Theaveragelengthof disabil-ityofthoseindividualsmissingatleast1dayofnormalactivitywas 49.19±77.56days.Thisstudyfoundthat34.2%ofinjuriesoccurred duringthedaytime,followedby28.1%inthemorning,23.5%during

sunsetand14.5%atnight.Additionally,43.9%ofinjuriesoccurred enroutetoworkorschool.

Fig.2demonstratestheanatomiclocationofinjuriesforall indi-viduals.Theyweresimilarforbothchildrenandadults,withlower extremityinjuriesrepresenting39.5%ofallinjuries.Injuriestothe upperextremitiesandheadrepresented16.3%and14.3%, respec-tively.Fracturesrepresented16.3%ofinjuries,andcorrelatedwith adisabilityofmorethan30days(25/57vs7/139,p=0.001).

Crash characteristicscomparingchildrenand adultsare rep-resented inTable 3.55% of allinjuries werepedestrianrelated (108/196).93%ofallchildreninvolvedinanRTIwereinvolvedasa pedestrian,while44.8%ofadultsinanRTIwerepedestrians(39/42 and69/154,p<0.001).Similarly,adultswereinjuredmoreas pas-sengers(85/154vs3/42,p<0.001).ThemajorityofRTIsoccurred onhighways,representing77.3%ofadultRTIsand50.0%of

child-Table3

Crashcharacteristicscomparedtochild/adulthood.

Childa _{Adult Total p}

Injuryseverity Minor 36 103 139 0.36 18.4% 52.6% 70.9% Major 6 51 57 0.83 3.1% 26.0% 29.1% Pedvspass. Pedestrian 39 69 108 <0.001 19.9% 35.2% 55.1% Passenger 3 85 88 <0.001 1.5% 43.4% 44.9% Commute

Goingto/fromwork/school 9 77 86 0.03

4.6% 39.3% 43.9% Goingelsewhere 33 77 110 0.09 16.9% 39.3% 56.1% Roadtype Highway 21 119 140 0.004 10.7% 60.7% 71.4% Pavednon-highway 2 16 18 0.38 1.0% 8.2% 9.2% Non-pavedroad 3 8 11 0.71 1.5% 4.1% 5.6%

Smallsidestreet 15 9 24 <0.001

7.7% 4.6% 12.2%

Parkinglot 2 2 n/a

1.0% 1.0% Playground 1 1 n/a 0.5% 0.5% N/A 11 30 41 0.54 5.6% 15.3% 20.9% Total 42_21.4% 154_78.6% 196

a_{Childisdefinedaslessthan15yearsold.}

hood RTIs(119/154 vs21/42, p=0.004). Notably,children were morelikelytobeinjuredonasmallunpavedsidestreetthanadults (15/42,vs9/154,p<0.001).

4. Discussion

ThisstudywascarriedouttodetermineRTIincidenceandcrash characteristicsinatestimoniallyhigh-riskareaofDaresSalaamin ordertodesignaninjurypreventionprogram.

Theresultsofthisstudyshowthatabout33outofevery1000 individuals reportedbeinginvolved ina RTIwithinthepast12 months.Thisstudysupportsothersources,whichhaveidentified RTIs as significantpublichealth problemin sub-Saharan Africa

(Andrewsetal.,1999;Asogwa,1992;Mocketal.,1999a,c;Moshiro

Fig.2.Overallinjuriessustained.

etal.,2001,2005;Oderoetal.,1997).Comparisontoother

popula-tionbasedstudiesisnotentirelyappropriate,sinceoftenthestudy areaisgeographicallywiderandcontainsanaverageofhigh-risk andlow-riskareas.Thisisparticularlythecasewhenexamining datapreviouslyavailablefromDaresSalaam.Forinstance,Moshiro etal.foundtheratetoonlybe5.98per1000person-years, con-siderablylowerthanthisstudy.Ofnote,thisstudyusedastricter definitionofRTIthanothers,sinceitincludedonlyindividualsthat missedat least1dayof activity,and applieda two-stage sam-plingtechnique,whichincludedtheentirecity(Kobusingyeetal.,

2001;Moshiroetal.,2001,2005).Whenadjustedtothat

defini-tion,ourcurrentstudyhadanadjustedincidencerateof24.2per 1000personyears.A2009studyfromNigeriafoundtheRTI inci-denceratetobe41.2per1000person-years,considerablyhigher, andaroundthesamelevel asboth aSriLankanand aUgandan population-basedstudywhich foundanincidencethereof 49.0 and38.9per1000person-years,respectively(Labinjoetal.,2009;

Moshiroetal.,2005;Navaratneetal.,2009).Usingtheexactsame

methodology,andfocusingexclusivelyonthetestimonially“worst” areaofAccra,a2009studybyAmend,whichiscurrentlyunder review,founda verysimilarrateof33.0 per1000person-years

(Guerreroetal.,2011).Whencomparedtocountrywideestimates

fromtheUnitedKingdomwheretheratewasfoundtobe4.3per 1000person years,thepointisclear,thatRTIsin thishigh-risk areaposeamajorpublichealthrisk(WorldHealthOrganization,

2009).

Thisstudyfoundthat4individualshadafatalRTIwithinthepast year.The“verbalautopsy”informationwascontradictorybetween familymembers,andsensational.Therehasbeenmuchwritten recentlyaboutthereliabilityof“verbalautopsy”(Yangetal.,2006). Ourfirsthandexperiencecontradictsthesepresumptions,though thenwasextremelylimited.ThereforethefatalRTIcharacteristics werenotincluded.Withregardstodeaths,otherauthorshavea reportedthat41%ofinjuryrelatedpre-hospitaldeathshavebeen attributedtoRTIsinDaresSalaam(Museruetal.,2002).

Pleasecitethisarticleinpressas:Zimmerman,K.,etal.,RoadtrafficinjuryincidenceandcrashcharacteristicsinDaresSalaam:Apopulation Individualsinjuredasapedestrianrepresentedoverhalfofall

theinjuries.Childrenwerefoundtobeinjuredpredominantlyas pedestrians,oftenonsmallunpavedsidestreets.Inastudywhich focusedonperceivedsusceptibilitytoRTIs,Astrømetal.(2006)

foundthat 78% ofthepeople includedin theirstudyin Dares Salaamperceivedbeinginjuredasapedestrianlikelyorverylikely. However,onehospital-basedstudydonebyMuseruetal.(2002)

foundthat67.8%ofparentsbelievedthat“accidentswere unpre-ventable”andoftenquotedtheSwahilisayingof“ajalihainakinga”, whichmakesfatethedeterminingfactorforRTIs.Itisclearthat pedestriansareparticularlyvulnerabletoRTIsandfuture preven-tionprogramsshouldfocusonthissusceptiblegroup.Basedonthe Astrømdata,apreventionprogramwouldbemosteffectiveifit focusedonthemessagethattheroadisaseriousthreatandthatan individualisabletominimizethatthreatbybeingproactiveabout safety(Astrømetal.,2006).

Understandingthetypeofroadsandmotorvehiclesinvolvedin RTIsisimportantinconstructingapreventionplan(Pedenetal., 2004).ThisstudyfoundthatmostRTIsoccuronahighwayandwas implicatedwith77.2%ofthoseindividualsmissingmorethan30 daysofactivity.ManyofthehighwaysandsidestreetsinDares Salaamdonothavesidepavementsforpedestriansorcyclistsand areoftenovercrowded(Museruetal.,2002).Additionally,thereare fewpedestriancrossingareas,whichshouldbeofparticular con-cerntoschoolsincloseproximitytohighways(Moshiroetal.,2005;

Museruetal.,2002).Since93%ofchildrenwereinvolvedinanRTI

asapedestrian,aprogramfocusingonthisRTIscenariois impor-tant.ManyoftheRTIsinvolvedindividualsbeinghitbyorinjured whileridinginacarormini-bustaxi.Assuch,preventionmeasures shouldtargetbothdriversandpassengersofthesevehicles.

WhereasotherstudieshavereportedthemajorityofRTIs occur-ringatsunset,thisstudyfoundthat34%ofinjuriesoccurredduring thedaytime,andaffectedindividualsduringtheirdailycommute

mostofthetime(Museruetal.,2002).Thismaybeaproductof

thestudyarea,sinceitrepresentsacommuterarea,dominatedby ahighway,andmayleadtoanoveremphasisonthisasaproblem inDaresSalaamingeneral.

Two-thirdsofthesurveyedpopulationtendedtoseekmedical attentionatahospital,andthreefourthsofthoseinvolvedinan RTIreceivedsomeformofroadsideassistance.Thisisinformation thatwe havenotseendescribedelsewhere.Additionally,police reportswerefiledforabouthalfofalltheinjuriessustained.These findingsmayhelpprovideawaytoestimatetheactualimpactof RTIsfromsecondarydata.Thatistosay,sincehalfofthe individ-ualsfiledpolicereports,perhapsadoublingofthepolicereport derivedsecondarydatamayprovideamoreaccurateestimationof theproblem.Thiswouldhavetobefurthervalidatedinordertobe usedconsistently,yetseemspromising.

5. Conclusion

Thisstudyencounteredlimitationsasaresultofcluster method-ology,theresearchenvironmentandasingle-stagesample.Ideally, ahighernumberdenominatoraswellasasamplingstrategyusing randomlyselectedindividuals,asopposedtogeographic cluster-ing,wouldhavebeenpreferred.Yet,theseidealpreferenceswould proveimpracticalinthisstudy’ssetting.Thisstudymodelleda 30-clustermethoddescribedbyHendersonandSundaresan(1982), whichoutlineshowtoachievealogicalbalancebetweentheideal and themore realistic optionsin a population-basedapproach. Althoughthisstudycouldnotcontrolforthelimitationsinherent toaclustersamplingstrategy,itdidaddressmany“classic” prob-lemsofgeographicclustersamplingduringthepilottestingphase. Theseincluded,teamspositioningthemselvesascloseaspossible toanindiscerniblelocation,andwhenfeasible,usingsidestreets

toavoidthe“mainstreet”bias.Thescopeofthisstudywasto iden-tifyspecificinformationforahigh-riskareawithinDaresSalaam. Therefore,whilesomeofthefindingsmaybegeneralizedtoother settings,theincidenceislikelytorepresentarate,whichishigher thanacity-wideaverage.

Inconclusion,thisstudyfoundthatRTIsinonehigh-riskarea ofDaresSalaamaccountforaconsiderableamountofdisability, resultinginamajorpublichealthburdenevidencedbyanaverage disabilityof49days.Childpedestrianswereaparticularly vulner-ablegroup.Injuriesresultinginafracturewerelikelytoresultina majordisability.RTIsinvolvingchildrenoccurredmostlyas pedes-triansandoftenonsmallunpavedsidestreets.Vulnerablegroups andspecificcrashscenariosshouldbeapriorityforinjury preven-tionprograms.HighwayswerethemostcommonsiteofRTIs,and thelowerextremitieswerefoundtobethemostcommonsiteof injury.PolicereportswerefiledforabouthalfofallreportedRTIsin thesetwowards,indicatingthatitmaybepossibletocorrector val-idatesecondarydata.Sincepopulationbaseddataisnowavailable inahigh-riskareaofDaresSalaam,atargetedinjuryprevention programmaynowbeconstructedandimplemented.

Fundingsource

The study was funded by Amend. However,all the authors workedonavoluntarybasis.

Competinginterestdeclaration

Therearenocompetingintereststodisclose. References

Andrews,C.N.,etal.,1999.RoadtrafficaccidentinjuriesinKampala.EastAfr.Med. J.76,189–194.

Asogwa,S.E.,1992.RoadtrafficaccidentsinNigeria:areviewandareappraisal. Accid.Anal.Prev.24(2),149–155.

Astrøm,A.N.,etal.,2006.Perceivedsusceptibilitytoandperceivedcausesofroad trafficinjuriesinanurbanandruralareaofTanzania.Accid.Anal.Prev.38,54–62. Bennett,S.,etal.,2002.Methodsfortheanalysisofincidenceratesincluster

ran-domizedtrials.Int.J.Epidemiol.31,839–846.

Chandran,A.,etal.,2010.Theburdenofunintentionalinjuriesandanagendafor progress.Epidemiol.Rev.32(1),110–120.

Dandona,R.,etal.,2008.Under-reportingofroadtrafficinjuriestothepolice:results fromtwodatasourcesinurbanIndia.Inj.Prev.14(December(6)),360–365. Dinh-Zarr,etal.,2001.Reviewsofevidenceregardinginterventionstoincreasethe

useofsafetybelts.Am.J.Prev.Med.21,48–65.

Guerrero,A.,etal.,PaediatricroadtrafficinjuriesinurbanGhana:apopulation-based study.Inj.Prev.,2011,doi:10.1136/ip.2010.028878.

Hayes,R.J.,Bennett,S.,1999.Simplesamplesizecalculationforcluster-randomized trials.Int.J.Epidemiol.28,319–326.

Henderson,R.H.,etal.,1973.Assessmentofvaccinationscarratesandsmallpox scarringinfiveareasofWestAfrica.Bull.WorldHealthOrgan.48(1),183–194. Henderson,R.H.,Sundaresan,T.,1982.Clustersamplingtoassessimmunization cov-erage:areviewofexperiencewithasimplifiedsamplingmethod.Bull.World HealthOrgan.60(2),253–260.

Kobusingye,O.,etal.,2001.InjurypatternsinruralandurbanUganda.Inj.Prev.7 (March(1)),46–50.

Labinjo,M.,etal.,2009.TheburdenofroadtrafficinjuriesinNigeria:resultsofa population-basedsurvey.Inj.Prev.15,157–162.

Macpherson,A.K.,Macarthur,C.,2002.Bicyclehelmetlegislation:evidencefor effec-tiveness.Pediatr.Res.52,472.

Mock,C.N.,etal.,1999a.IncidenceandoutcomeofinjuryinGhana:a community-basedsurvey.Bull.WorldHealthOrgan.77(12),955–964.

Mock,C.N.,etal.,1999b.TheeffectofrecallonincidentratesforinjuriesinGhana. Int.J.Epidemiol.28,750–755.

Mock,C.N.,etal.,1999c.Epidemiologyoftransportation-relatedinjuriesinGhana. Accid.Anal.Prev.31,359–370.

Moshiro,C.,etal.,2001.Theimportanceofinjuryasacauseofdeathinsub-Saharan Africa:resultsofacommunity-basedstudyinTanzania.PublicHealth115, 96–102.

Moshiro,C.,etal.,2005.InjurymorbidityinanurbanandruralareaofTanzania:an epidemiologicalsurvey.BMCPublicHealth28(5),11–21.

Museru,L.M.,etal.,2002.Patternsofroadtrafficinjuriesandassociatedfactors amongschool-agedchildreninDaresSalaam.Afr.Saf.Promot.1,37–43. Navaratne,K.V.,etal.,2009.Population-basedestimatesofinjuriesinSriLanka.Inj.

Odero,W.,etal.,1997.Roadtrafficinjuriesindevelopingcountries:acomprehensive reviewofepidemiologicalstudies.Trop.Med.Int.Health2,445–460. Peden,M.,etal.,2004.WorldReportonRoadTrafficInjuryPrevention.WHO

Pub-lishing,pp.4–7.

Peden,M.,etal.,2009.WorldReportonChildInjuryPrevention.WHOPublishing, pp.31–35.

Redelmeier,D.A.,etal.,2003.Traffic-lawenforcementandriskofdeathfrom motor-vehiclecrashes:case-crossoverstudy.Lancet361,2177–2182.

Servadei,F.,etal.,2003.EffectsofItaly’smotorcyclehelmetlawontraumaticbrain injuries.Inj.Prev.9,257–260.

Shults,R.A.,etal.,2001.Reviewsofevidenceregardinginterventionstoreduce alcohol-impaireddriving.Am.J.Prev.Med.21,66–88.

SpinksA,etal.,2005.TheWHOsafecommunitiesmodelforthepreventionofinjuries inwholepopulations.CochraneDatabaseSyst.Rev.18(2).

WorldHealthOrganization,2009.Globalstatusreportonroadsafety:timeforaction, www.who.int/violenceinjuryprevention/roadsafetystatus/2009.

Yang,G.,etal.,2006.Validationofverbalautopsyproceduresforadultdeathsin China.Int.J.Epidemiol.35(3),748–750.