Spatio - Temporal Analysis of Characteristics and Causes of Road Traffic Crashes in Oyo State of Nigeria
*1Grace Oluwatoyin Korter, *Olusanya Elisa Olubusoye, 'Mees Adebare Salisu
pepartmen~s of *S~at~stics and #Economics, University of Ibadan and Department of Mathematws & Stat~stws, Federal Polytechnic, Offa, Nigeria.
ABSTRACT Keywords: Spatial Hotspots; Road Traffic Crashes; Injuries and Deaths; Nigeria. BACKGROUND Millions of people each year will spend long Correspondence: G. 0. Korter weeks in hospital after severe crashes. The WHO Email: [email protected] estimated that 1.3million people were killed by road traffic crashes (RTC) and 50 million INTRODUCTION injured on the worlds road annually, adding Millions of people each year will spend long weeks that over 80 percent of the figure occurred in in hospital after severe crashes and many will developing countries, with Africa having the never be able to live, work or play as they used to highest death rate. Also, WHO predicted that if do. In some instances, a portion of persons injured nothing is done by countries to stem the tide end up dying. Road traffic injuries (RTI) are death by RTC would increase by 65% by 2015 t~ growing health issues disproportionately affecting 2020, overtaking malaria and tuberculosis. vulnerable groups ofroad users, including the poor. This paper is aimed at investigating the More than half the people killed in traffic crashes characteristics, causes and spill over effects of are young adults aged between 15 and 44years RTC in Oyo state for two periods, namely years often the bread winners in a family. The World 2011 and2012. Health Organization (WHO)[l] concludes thatRTI cost low income and middle income countries METHODS between 1 percent and 2 percent of their gross The total number of RTC cases and causes national product more than the total development recorded in Oyo state where observed for the two aid received by these countries. periods under study. The total number of persons killed, number injured, sex and age of Furthermore, the WHO estimated that 1.3million victims on each RTC incident was obtained. The people were killed by road traffic crashes (RTC) causes were classified under four headings: and 50 million injured on the worlds road annually dangerous driving, speed limit violation . ' adding that over 80 percent of the figure occurred mechamcal fault and human factors. The in developing countries, with Mrica having the multiple bar chart was used for comparative highest death rate. Also, WHO predicted that if purposes. The Moran's Index and a nothing is done by countries to stem the tide, death complimenting statistic; the Getis and Ord by RTC would increase by 65% from 2015 to 2020 statistic was used to ascertain spill over effects. overtaking malaria and tuberculosis [2]. ' RESULT In 2010 the governments of the world declared Number of RTC varied over the two periods. 2011-2020 as the Decade of Action for Road Safety. RTC is characterized by deaths and injuries of The unanimous support for this Decade of Action adult males. The causes vary from one LGA to from Member States explains the growing another, but similar within contiguous LGAs. awareness of the devastating scale of RTI as a global public health and development problem. CONCLUSION RTI are estimated to be the eighth leading cause of The results should enable the orientation of death globally. They are the leading cause of death deaths and injury prevention policies targeted for young people aged 15-29 years, and as a result on the adult males in the state. take a heavy toll on those entering their most productive years. Economically disadvantaged families are hardest hit by both direct medical
The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 1Pagell1@ Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A. costs and indirect costs such as lost wages that in Belgium. The paper analysed the geographical result from these injuries. Despite the enormous trends over a 20 year period at the district level. toll exacted by RTI, RTI has for many years been The geographical analysis showed marked neglected by global health and development differences between districts. Even though a agendas. This is regardless of the fact that RTI are favourable trend was observed for the traffic largely preventable and that the evidence base for injuries deaths in Belgium it is important to effective interventions is extensive [3]. highlight the important slowing down of this trend during the most recent years. The study concluded The occurrence of road accidents inflicts burden on it is also necessary to underline the importance of the community which may be considered in two geographical disparities in the distribution of parts. Firstly are the pain, fear and suffering YPLL rates within the entire population. imposed by the occurrence, or the risk of occurrence of road accidents. These are considered Cirera et al [7] described the characteristics of of great importance in a society that values human motor vehicle (MV) injury cases admitted to life and human welfare. Second, is the more Emergency departments (ED), and assessed concrete and ascertainable burdens in the form of factors related to injury severity and hospital the net loss of output of goods and services due to admission. The subjects were MV injury patients, death and injury and the expenditure of resources aged 16 or more and admitted to four EDs in the necessary to make good the effects of accidents. city of Barcelona (Spain) from July 1995 to June 1996. Severity was assessed with the Abbreviated Reynolds [4] in a study in Britain estimated the Injury Scale and the Injury Severity Score. burdens of medical expenses, vehicle repairs and Univariate and bivariate descriptive statistical costs of administration of RTI. The study analyses were performed, as well as multiple attempted to measure how much the community of logistic regressions. For the 3791 MV injury cases Great Britain is worse off in terms of measurable included in the study period, a larger contribution costs by the existence of road accidents. The of cases was noted for males (63.1 %), for cases estimates contained only the minimum values that younger than 30 years (55.3%) and for motorcycle can be placed on the consequence of road accidents, or moped occupants (47.1%). Mter adjusting for which cannot be a complete guide to the policy to be age, sex and the presence of multiple injuries, adopted and the expenditure to be incurred on pedestrians, followed by moped and motorcycle accident prevention. The effects of road accidents occupants were at a higher risk of a more severe were discussed summarily as damage to property, injury. Correspondingly, these user groups also medical costs, administrative costs, net reduction showed a higher likelihood of a hospital admission in output of goods and services, with allowance when attended to in an ED. Injury cases attended being made in the case of persons killed and to in the ED during night hours were also at a economic effects of a smaller population and a higher risk of a hospital admission. slightly different age/sex occupational structure. Labinjo et al [8] explored the epidemiology of RTI Plasencia and Borrell [5] estimated the incidence in Nigeria and provided data on the populations of morbidity and mortality due to injuries in the affected and risk factors for RTI. The RTI rates for population over the age of 14 years in Barcelona, rural and urban respondents were not Spain. Injury distribution according to sex, age, significantly different. Increased risk of injury was external cause, place of occurrence of the injury associated with male gender among those aged 18 and severity was also obtained. Injury morbidity to 44 years. Simple extrapolations from the survey and mortality amongst residents of Barcelona suggested that over 4 million people may be follow sex, age and cause of injury patterns which injured and as many as 200,000 potentially killed are, overall, comparable to those observed in other as the result ofthe menace annually. industrialized countries, suggesting that similar etiologic factors might be operating in those areas. In order to enhance prioritization of the burden of RTC, Chandran et al [9] calculated years oflife lost RTI are a major public health problem. Leveque et and reduction in life expectancy using population al [6] chose Years of Potential Life Lost (YPLL) to and crash data from Brazil's ministries of health analyse the trends during the period 1974 to 1994 and transport. The potential for reduction in crash and the relative impact of the traffic injuries death mortality was calculated for hypothetical on total mortality and on total avoidable mortality scenarios reducing death rates to those of the best
The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPagellfj Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A. performing region and age category. For males, at factors that have helped to check the frequency of birth, RTC reduced the life expectancy by 0.8 years road mishaps and indirectly led to a decline in the and 0.2 years for the females. The study concluded trend ofroad traffic crashes from 1988[11]. many years oflife lost for men and women could be averted if all rates matched those of the lowest risk In Nigeria, RTC death rate is 162 deaths per region and age category. 100,000 populations as stated by Ogbodo and Nduoma[12]. This is against the expected world Ezel et al [10] piloted a systematic mortuary-based average of 22 deaths per 100,000 populations data collection in Ibadan to determine the nature Sukhai et al [13]. Thus, the Nigerian RTC death and circumstances of fatal RTI and assess data rate is disproportionately high in comparison to quality against existing data sources. Using a draft the world average by over 636%. This paper is data collection system developed jointly by WHO aimed at investigating the characteristics, causes and Monash University, the detailed information and spill over effects of RTC in Oyo state, Nigeria was prospectively collected on RTI University for two periods, namely years 2011 and 2012. The College Hospital mortuary admissions in Ibadan state is the major link between the northern parts from September 2010 to February 2011. of the country and the busiest city (Lagos). The Demographics, road user type, counterpart focus is on the frequency of deaths and injuries that vehicle, intent, manner and medical cause of death results from this menace and thus will enable the were recorded. Findings showed mortuary orientation of deaths and injury prevention admissions included 80 fatal RTI cases: 81.3% policies in the state. males. By road user category, 28 (35.0%) were pedestrians; 28 (35.0%) motorised 2 wheeler users; The research question is whether there is spill over 18.8% car occupants; and 11.3% bus occupants. In effects (spatial dependence) of causes of RTC 70% of cases, medical cause of death was head across the study area. That is, for instance, does injury, including 25 of 28 motorised 2 wheeler the cause ofRTC at a particular location influence users (89.3%). Estimates from this study indicate the cause at a neighbouring locality? The apparent increased mortuary capture of fatal RTI underlying assumption is that RTC are caused by compared with police data. The paper certain actions of man, which may not necessarily demonstrates the feasibility of collecting detailed, be aimed at causing accidents. These human timely RTI fatality data through mortuary based behaviours invariably, make RTC unavoidable. surveillance in Ibadan. While not all RTI deaths are reported to any authority in Ibadan, this large METHODS case series compliments existing data sources and Data and Data Source suggests that pedestrians and motorised 2- Data on all recorded RTC cases and traffic volume wheeler users die most often in RTC. Frequent for 2011 and 2012 was obtained from the Federal head injuries among motorised 2-wheeler users Road Safety Commission RS11.3 Oyo sector strongly support the need for helmet wearing. command with headquarters at Eleyele, Ibadan in Oyo state. The FRSC Oyo sector command In February 1988, the Federal Government comprise of ten unit commands, each unit created the Federal Road Safety Commission command has designated service routes within the through Decree No. 45 referred to in the statute Local Government Areas (LGAs). The unit books as the FRSC Act cap 141 Laws of the commands and the LGA they oversee are as Federation ofNigeria (LFN). Prior to this time, the follows: RS11.30 Eleyele unit command: Ibadan Nigerian Police Force established in 1930 was North West and Ibadan South West, RS11.31 saddled with the responsibility of keeping road Ogbomoso unit command: Ogbomoso South, traffic accident records. The Federal Government's Ogbomoso North, Ogo Oluwa and Surulere, intervention on road traffic accidents by the RS11.32 Oluyole unit command: Ibadan South establishment of FRSC in 1988 led to concrete, East, Oluyole, Ona Ara, RS 11.33 Iddo unit deliberate and sustained policy action to address command: Ibarapa North, Ibarapa Central, road safety issues. Activities of FRSC such as Ibarapa, East and Ido, RS11. 34 Mokola unit overhauling the national drivers' license scheme, command: Ibadan North and Ibadan North East, review and re-engineering of the vehicle license RS 11.35 Egbeda unit command: Egbeda and data base, highway patrol, ambulance services, Lagelu, RS 11.36 Saki unit command: Saki West, public education programme on drivers' road Saki East, Orelope, Atisbo, Iwajowa, Kajola, traffic habits and keeping of offenders registry are Itesiwaju, RS11.37 Kisi unit command: Irepo,
The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPage liM Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A.
Olorunsogo and Orire, RS11.38 Atiba unit The Moran's I statistic is structured as a Pearson command: Atiba, Afijio, Oyo East, Oyo West, Iseyin product moment correlation coefficient, plus W, the and RS 11.39 Moniya unit command oversees only contiguity weights matrix. Y is a covariance Akinyele LGA. matrix, that is, the relation between the spatial units is c~culated as(yi _ YXl· _y )- The obtained The unit command, road/route, location and time 1 measure IS scaled by of RTC were obtained. The type of vehicles that were involved in the crashes was reported as either 2 government diplomat, commercial or private n ~ [t(yi- Y) ] vehicles. The vehicles' registration numbers, the LLWij i=1 number of passengers, cause of RTC and total i=1 j=1 casualties namely; the number of adults injured, By convention, i * j . As a result, number of children injured and the gender of injured persons. Also, the number of adults killed, number of children killed and the gender of 1 = __n__ _i=_1....: 1'-·=_1 ______persons killed were included in the RTC data. The n n possible cause of each incidence of accidents is LLW!i i=1 j=1 recorded along other information. ,wherey, =the value ofvariabley on segment i,y = The possible causes of accident reported were the mean of variable y, n =the number of segments, grouped into four categories for the purpose of this w u = a weight indicating if segment i is connected to study as follows. Dangerous Driving (DVD): segmentj (e.g. 1) orifit is not (e.g.O). dangerous driving, wrongful overtaking and dangerous overtaking; Speed Limit Violation Gi (d) Statistic (SPV): speed limit violation and loss of control; This statistic developed by Getis and Ord [15] in Mechanical Fault (MF): brake failure and tyre 1992 identifies hotspots and measures the degree burst and Human Factors (HF): driving under of association (spatial dependence) that results alcohol influence, overloading, obstruction and bad from the concentration of weighted points (or area road. represented by a weighted point) and all other Analysis weighted points included within a radius of distance r. = " W..Y.. ' z' -+ j' Descriptive statistics and geographic information I £...Jj lJ lJ -t- system approach was adopted. Multiple bar charts were used for description and comparative We assume an area subdivided into n regions, purposes, while the Moran's Index and a i = 1,2, ... , n, where each region is identified with a complimenting statistic (Getis and Ord) were used point whose Cartesian coordinates are known. for hotspot analysis and spatial dependence (spill Each i has associated with it a value y (a weight) over effects) assessment. taken from variable Y. The variable has a natural
origin and is positive. n Moran's I Statistic LWi.(d)y.1 Is a global measure for hotspots and spatial The statistic is, Gld )= ·=1 n !I ,j not equal to dependence developed by Moran [14] in 1948. The index measures spatial dependence based on LYj j=1 feature locations and attribute values. The measure evaluates whether the pattern is i , where ( w ij ) is a symmetric one/zero spatial clustered, dispersed or random. The null weight matrix with ones for all links defined as hypothesis states that the feature values are being within distance d of a given i; all other links randomly distributed across the study area. When are zero including the link of point i to itself. The the z score or p value indicates statistical numerator is the sum of ally1 within d of i but not significance, a positive Moran's I index value including y 1• The denominator is the sum of all y 1 indicates tendency towards clustering while a notincludingy1• negative Moran's I index value indicates tendency toward dispersion.
The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPagel@) Spatlo-Temporal Analysis or Road Trame Cl'aahes- Korter o.o. au~ o.E, Sallsu A.A.
RBSVLT8 Next, we colllliderthe catego:riea ofperBGII.B ~ured. CharaoteristtN in RTC. Figures 3 shows th& number of males Figure 1 shows number of RTC caaea in (blue) and femalell (yellow) ill,junld in 2Gll. 20ll(yeUow) and 2012 (blue). Figure .2 llhow11 the J:i'ipres 4 shows the number of malei {bl\l.e) and total casuelties in2011(yellow) and 2012 (blue). females (yeUow)iJVaredin2012.
.._.. .lJ ...... + .... _,... Vlf::.U .-) 0:0 .. ------ Fica.re 8: Geovisual presentation of --- -- . "' -. Ma/Q/Feroolu I~ in Road 'I'ro/'fk A~ 2011 N N i + t Leg4!nd -] ""- + ~-~.. -11)•_?..0,, t> ., 2{1 ..:. 00 ~."l ------.. - "" -----. Ficure -4:: Geovisual presentation of Ma/Q/Fertl4lu I~ in Road 'I'ro/'fkAccidents 2012 Tho Nigerian Heelltl Joumel, Vol. 13, No 4, october-December, 2013 IPaoalfi.) SpaUo- Temporal Ana1y1k of Road Trafllc C!&.ahaa-Karter G.O, Olubusoye O.E, Sall9u A.A. Figure 5 shows the nlli.Dber of adult. (blue) anJl Figure 7 shows the number of malee (bl11e) and childnm (yellow)that BUBtained~urisa u a teiiUlt femalee (yellow) killed in 2011. Figure 8 shows the of RTC in 2011. Figure 8 ahows the number of nomber ofmales (blue) and females (yellow) lrJ.Ihld adults (blue) and children (yellow) that IIWitained in2012. ~uriea 11.1 a retlult ofRTC in 2012. ... • • .. .10 .., • .._,..,._ .... Fltrare 7: Oeovisual presentation of Picure 1: Oeovisual presentation of Mala IFe1714kll killed Rood 7'roffic Al:dcleAtB Ch.ildrm!Adv.zt.IT+JuredmRoad'l'rofficAI:ci.tl.enle m 2011 2012 N -·- t \ ··~· ;- '- · -~.. ,__, __ . '...... l : • ••-4 ,r' ... - I ] ·-- - ~ -·~ .. \ 1 .,.. -· ...., + - l.=) ...... -:t:. .- ..;. ·- ... c~~" - . •»~ - - - Figure 8: Geo11 ieual prnentatJon of Mala/Fem.al.es killed ill Road 7'roffic ADcidm.b J'igure 8: Oeovi.uol preeentotion of 2012 Ch.ildrm 1AduZu IT+fruwd mRood 'l'rofficAI:ci.tl.enle 2011 Figure9showsthenumberofcbildnm(yellow)and adults (blue) killecl in 2011. Figure 10 llhowa the In t.ha fOllowing OO!!•idlll' aacti011, we the cataguri.aa numbel'ofc:hil..dl'en(yellow)anJladulta(blue)ldllocl ofpen~OIIIl inRTC. killed in2012. The Nigerian Healltl Joumel. Vol. 13, No 4, October-~. 2013 IPaQoiiil Spatlo-TempomiAnalysls of Road 'n'afllc CITIIIhes- KDI1er G.O, Ofubuaoye O.E, SallsuAA DaacerouaDrivhqJ SpiD. overetr~(Moran) 'l'he nuD hypothe!l'is states that the spatial pattem israndom acroaa thestuily area. In 2011, the Moran'a I Index calculated wu 0.29, while the Z ~~core wu 3.61 standard deviatiOIIJI, with a P value ot 0.01. For 2012, the Moran's I ID.da. cBJculakd wu 0.86, while the Z acore was 4.37 standard deviation.e, with a P valu ot 0.01. Thia indicatea leaa than 1% ljkeHbood that this clustered pattern could be the result of random chance.(seeal.ao, Table 1). N ..- < J ...... , •• . " -- i "" ,---. Figure 9; Geovisuc;l pruentc;#.on of '·' --- . . '·. Cltild.tm/.Arl.ull811illed in &ad »a/fit ~na i ' ;. _) - 2011 l.., - ·-·· r' --·--· N _~·__ ..... _ _Q~--h··-__ ..,- ___·.. ------·-.. -_ 1:""::":1 •n·•..,..,,,~ ...... _.._,...... , ., i _,_...... -~- Figure ll; &lo!>Urual p~n of lwtttpots for ~uo~Drivmg in 2011 (MQI"QQ{s) ,. \ "•4-.o• --.... ., I ~ ' ' .... l., - - -f ( .... _._ ... _.. _ Figure 10; Geouisuc;l presentc;tion of ..... ___ .. ...:.. •. -.s ...... C1t.iltl.mt.I.Arl.ull8 •itl«li., Road 'I'I'afflc .Ar!Citkna ~-- ...... - -~ --··- .... -·-·J•- ·---...... ~··· 2012 - -...... ·· ,.,.-.. G Ml ~ • .:.;=::.-· c.- Th& spatial dependence V8lu.es usociated. with causea ofRTC aregiven in thenextaection. Pigure U: &lo!>UvaZ p~n ofhotllpots for ~uoll3Drivi116in 2Ql2(MQI"QQ{s) 111e Nigerian Heelltl Joumel, Vol. 13, No 4, Odlober-December, 2tl13 IP8flalf,j Spalfo-T~ Analysis of Road Trafll~ C!a&h811- Kor11Br G.O, Olubu90)'9 O.E, Sallsu A.A. HoA&pow(M'onm) 'l'h& h.otapo~ u a rtSl.llt of dangerous driving in 2011 (lljgunlll) were on uuijor toada within Oyo Weet, Oyo East, Afijio and Ibarapa East LGA3. Whilb, the hotspots as a resl.llt of dangerous #_,.,. \ ...._ _ --··\ driving in 2012 (Figure 12) 'W'e'N on ms,ior :roads I ,.,..,. l _, within.Atiba, Oyo East, Oyo Westand.Afijio LGAa. 1 Table l: Momn'11 Spemal DepenckMe for ... .-.... \ ~ Drivi.ne o.mtmg LoctJ. ~nt ---_) Aml6 (LGA): Number ofDcnge:ro!UI Driving in 0,.0 t "".. . .) - St4k,fd()11-2012 t I J / ...... _H. - 'DU.JIIU liU,liOIS t""'"''r"1, Lll!loaa ( -...,...... ' ua I'M' • LM! Ptllot o.-ollaa LC.\ Lll!loaa ... __.. Utili • UiPnma C illb l The Nlgellln Heallll Joumel, Vol. 13, No 4, October-~. 2013 1Pagelif1 Spatlo-TempomiAnalysls of Road 'n'afllc CITIIIhes- KDI1er G.O, Ofubuaoye O.E, SallsuAA Speed LimitViolatiDn Houpote (Moran) Spill over eft'ecta (Moran) The hotspots aa a nllllllt ofspeed limit violation in The null Jcypotb.e.sia Bt.Ues that the llp8tial pattem 2011 (Fig\lre 16) were on ~ re~ within i& :random acro.~~~J the 8tody area. Oluyole, Ona Ara and Ibadu Scraib. East LGAs. In 2011, the Mo:ran'sl Indez. calculated equal 0.14. While, the hotspotJJ as a remt of speed limit The Z score equal1.95 standard deviations, while violation in 2012 (lligunl16) were on ms,ior roada thePvalueequal0.10.1'hereisleBBthanot.o 1~ within Oluyole, Ona Ara and lbed•n South East likelihood that this clustenld pattern could be a LGAs. 'rel!Ult of:random ehanoe. For 2012, the Moran's I Indexcalc11lated equal 0.17. TheZ score equal2.38, Table 8: Moron!• Sp I I_ ....~ - \ ... _ ,_,.... I ' ' ( ' j ' ' "- ·•· B'sy: HH =HiPHigh; mSE= lbadanSouthEaat "' I ...... T 11·1• • ' Reilult reveals high high values ofconcentration of t I I ~~ _ ...... 1'-,..d'"t RTC clustering within Oluyole,Ona Ara and. lbadanSouthEuti..GAI! for2011 and2012. ~_;:::_.._,.?"" ...... ·-- . -··· ...... _ SpiD. over effilc:ta {Gt!ti.s andOrd) ... - :: .'!'~ ::- The nuD. hypotheeie etatee that the llp8tial pattem ...... ~.. - israndoma.aosathestuilyarea.. The mrulw &om the G atati.stic frlr teeting spatial :Fiproe 1& &!cvi8uol p~n of~ fw dependence are given aa fullow: In 2011. the G Sp«dLimit Violation in .2011 (Moron's) Index was 0.29 with aZ score of2.55. There was a Ieee than 6 ~likelihood that the clu8teriDg ofhigh values is the niSult of random chance. While, for 2012, the G Wll8 equal 0.81 while the Z ec:ore equlled 3.01 standard deviations. There was a lee a than 1'Ill JjlreJjbood that the eluatering ofhigh ', - ' value8 could be the l'ejjult of random ehance. < '· - --... ~.. al.aoTahle4) 1 --..- ·-· --- ·...... •' • ...... -.;;-..... - j ...... ~-- r ...... - ~.. - ...... Uo'lt ~...... "_..., ..._ .. --,~ .... _ ..... _ • ...,__...... - .. 4 _ ~- - ::·!:.~: ----, ... ' ...... = :.: ·~::::~:: 0 '~~' lO .-; «' .,"'__ - · ·;~:.:~:~::· ... ..t.,..-.- .... -··.. -· - ··--~- J'fpre 1& 111e Nigerian Heeltll Joumel, Vol. 13, No 4, Odlober-December, 21113 IP8flalfll Spatio-Tempo rei Analytit of Ro.c1 Traflie CrashM-K«ter G.O, Olu~ O.E, 8elitu A.A. The null h;ypoth811ia statea that the spatial patklm ill randomamK!Bthe Btud,yarea. N ~ In 2011, the Moran's I Index calC'IIlated waa 0.28, • with a Z ecor:e of 3.52, and P value of 0.01. Thill __. .... ,j. - --· ! .. \ ... ~... :.- .- -... ~ indicataa leas than 1%> likelihood that this .. -. ' .- ' cluater:ed pattern cwld be the reeult of random . '· chance, For 2012, the Moran's I Inclel: calculated Wllll 0.4:7, whiletheZ IICOl'e Wllll 6.54, with a Pvalue ot 0.01 and a less than 1<;{, liblihood that thi.8 cluater:ed pattern cwld be the reeult of random dumoe(aeealao, Table6). N ...... - i '·· Jr.t&ure 18: b160D£9UIIl p~ Of!Uit.8pot8 /CIT' Speed Limit V!Okdion in 2lJ12 (GeliB and Ord) \. ' . ' ·-_ r ·-· \ o'Zf_=-..;.:;:,~- / - ... ' "'..... ;- ~ ., ...... - \ ) Hotspots(GetiaandOrd} -...... - .r • 1 . . y · ~· .:; The hotspota tor RTC as a reault of speed limit .' . ; ...... "·-··-. (Figure ~r <''_-{_,.. ~ - violation in 2011 17) were on roads ···--.:· .,._;" within Olu,yole LGA. The con.centrati.on is leas ....o;o •.•._.,._...... mgnificant in Ona Ara, Ibadan North, Ibadan ~:.~~;.:: - =IIi.- ...... ,..,...... North East, lbadan North West, lbadan South ~ ...... h. East, and !baden South West LGAs. In 2012 the hotpots ar:ethe aame with the later(Bee F5gure 18). Table 4: lhtis and Ord Spatial ~ for ~ 191 Geovimtd p1"e8m11Jtion. uf hotllpot8 for Speed V!Okdion CUMTIC Local Limit Govemnumt M~Fcudt in. 2011 (.Mornn'11) Area. (!.GA.~): Numhu o{Speed Limit Viola#o~ in OyoSto.ts, 2()11-2012 N (lllGBEBT POBI7'1.VE VA.l.UBB) ...... - i , -~, \ ...... --·!' _;' :.. - ·- -- ) I I - _.._ I .... 1'' ..,...... , ..... '-"1111<_ . ...<1•.. ffff ..- -~ ...... ,...~... - -h_---, ....• ...... _,... .. The hotl!pots inelude Oluyole, Ona Ara and _ ...... c II' ';11(1 .n In .!W)... .. _ _ EgbedaLGA.sin2011and2012. _...... Mechanical Fault ----- __. _------·r· -·--'" ·--·---·- -~ -----r---~-- Spilloverelmcts (Moran) Medlonical Fault in 2012 (Mo~s) The Nigerian HHith Joumal, Vol. 13, No 4, Odr..,., December. 2013 IPageCij.t Spalfo-T~ Analyslll of Road Traffic; Clashes-Kor11Br G.O, Olubu90)'9 O.E, Sallsu AA Hotspots (Moran) Thehotapota as a result of' mechanical fault in 2011 were on major roade within Oluyole, Ona Ar:a, N I..agelu, lbadan South East and Egbeda LGAs (Figure 19). In addition to theae LGAs, in 201.2, I ' ) ....,. ,. Lagelu LGA became more concentrated with i accidents 1"88ulting from mechanical fault {aee Figure.20). ·' ...... Table 6: Moran'~ SpotWl Dtptntktlff for MechcuticcU Fa:ult DmiJ118 Local Govtmm4nt Anu18 a.GA>: Number t1{ MechcuticcU Fault mO,yo St4k, 20D-2012 ....t•c•._ • • f ...... ··-_,...... ; .. """ """~'~'"; ':___] ...... : .... .- ·u...... ,. ,"~...... • • • u ... 11• • Key: HH =High High; mSE= DJadRll SouthEast Reault reveal$ high high val~ ofconcentration of N RTC clusk!ring due to mechanical !aalt within Oluyole, lbadan South East. Ona Ar:a, Egbeda and "... .. - i I..qelu LGAsfor2011 and2012. - ' \ .:.... - ,..J ..I Ho" •~<-•-•· \, Spilloveret'feet.11(GiltisandO:rd) r ,- ·. \ The null h.Jpothea:ie statee thatthe spatial pattern \ n.co- I / \ ___...-(--· .... is random aero liB theatudy area. _) -... The re9ults from th& G statistic for testing spatial - /... -P...· , ....~ .. { dependence are given as follow: fur 2011, the G ll) Index was 0.84, while the Z IICOI'e was 8.91 - ' r ... """' ...- ~\·-.··--+, standard deviations with leiiB than 1 % liblihood '"-·::... r that the clu.stciDg of' high values coa1d be the u~~ ·~...... re9ult of r8lldom chaooe While, for 2012, the G (.,0:-· ...... , ...... _ .. index was 0.5 while the Z score equalled 5.68 ... ·~··.. =., ... _ ·-"-· llta:n.dard deviations with less than 1% Jikelibood :::::::::; ~':..:::.'' ;:." ~ Hotspota(GetisandOrd) Fipre i2: Gewisual p~ t1{hatttpo#B for The hotapota for RTC as a reault of mechanical MeclumicalF4Ull in. 2012(0emG11d()rd) fault in 2011 (Figure 21) were on me,jor road8 within Egbeda. Ona Ar:a, Oluyole, Ibadan South Ta'ble 8: G!ti8 Gild Ord Spatitll Dttpttn.tkn.ce for East, DJadRll South West, Ibadan North Eaat, MI!Manical Fault Clml:1ll(l Local Govem.ment.AnM lbadan North Weet and Ibadan North LGAs. 111& (LGAJJ): Numhu of Rood Traffic Accide~ hotapota in 2012 wen~ the aame LGAs (see Figure Ruultin.g from MecJwnical Fmilta in Oyo Staat. 22). 2011-2012 'T'he Nlgellln Health Joumel, Vol. 13, No 4, October-~. 2013 IPageiil.J Spalfo-T~ Analyslll of Road Traffic; Clashes-Kor11Br G.O, Olubu90)'9 O.E, Sallsu AA (HIGIIBS7' POBITIVB VALVBB) lllai.!Oil n:AB.lllll! O,Pnla LOA o,za.... I~I'..U.. ~za.... U'lf 0.101 ....uea OJI:II .... O.Aoa il.l!e JIQI ON .... Ulll 0.000 • om;,.lo ~0118 ...... Mot ~- 11.018 -Nd._ ---z.e«)...... --- -· ...... ! . ll.OOII ( ...,.. -r- UIIT O.(lll I ' ---U«) Q.OI8 ::,mNrila.tll.iiO...,... --- ' Gl.ll:l llb<,l1lo CUIIll --- Ulll ---U«) Q.006 ...._ 11M -NIIIIO US( CUIOI -~~· ~_...... - _ --- ~ ...... -n.. For the two periods under consid81'ation the - -iiiiii...... ::.::::.:::·- ...... __.. ~ conce:ntntion of :znnnber of RTC attzibutable to _.... -.... mechanical fault i$ high in Egb&da, Ona Ara, Oluyole, Ibadan South Eaat, Ibadan South Weat, lbadan North East, lhadiiil North West, and Jbadan North LGAs. :Ficue 14: Geovieual p~n. of iwl8pot8 for Humon. Ftldor in 2012 (.MOTG.n'8) Spill OV81' efl'ed.a (Moran) Human Factor Hotepots (Moran) The null hypotheais states thatthe spatial patWrn The hotspots au a reuult ofhuman factors in 2011 is random ac:!'CIBII theatudy ll:nla. were onm.lijorroad.swithin Oyo Weatand Oyo East In 2011, the Moraa.'11 I Indsx cakulated equalled LGAII (Fi(w.'e 23). In 2012, the concentration. 0.18. The Z score equalled 2.28 &tandard becamehigherinOluyole,OnaAra,IbadanNorth, deviations, while the P 'Vlllue Willi 0.05, with leaa ThadanN~Eu~IbadanN~W~~Thadan than 6% likelihood that thi$ c}l,ljjtered pattern So~;~th Ea&t, and lbadan So"Oth West LGAll (Figure couldbethereeultofrandomch.anca 24). For 2012, the Moran's I Index. calculated. Willi 0.7, with a z scora of8.02 standard deviations, and a P Table 'l;Moran'tt Sp<#KU ~for Hu111«1f. 'Vlllue of 0.01 indicating a leus than 1% Jiblihood FGC~rJrs among Local~ Analt: Numh I \ 1 ...... --- '"'•t-• > I - . ·-· I) -' ( ...... ,.., .. ~ f"" ...... ,, ...... a ...... -- ... ~,~~,,~...... _ ...... _ Result reveal~~ hip high val~;~es ofcon.oentration of =:i '*" ...... ~ 0 , .. a; accident cluaterin&' due to human fil.ctore within --.....· ··--, ~ ··...... Ola,yole, Ona Ara, Ibadan South Eu~ Thadan = So~;~th West, lhedan North Eaat, lbadan North Pfeure 281 G»ois1u:JZ p1'U:Iln.tGticn of hotspots for Weat,lbadanNorthandEgbedaLGAain2012• .Por HunumFactor in 2011 (Mornn'8) 2011, the conoentntion wu low in. Oyo West and OyoEastLGA. 'T'he Nlgellln Health Joumel, Vol. 13, No 4, October-~. 2013 Spatio-Tempo rei Analytit of Ro.c1 Traflie CrashM-K«ter G.O. ~~ O.E, 8elilu AA Spilloveretl'ecta(GetisandOrd} Table 8; Chew Wid OnlSpat:W ~for The nallhypothesi& etate. that the eplill:i.al pattern Human Fact41'8 among r-1 Goventmellt .AmJ.: i.enmdomac:t'l)88th&&tud,yarea. NumbuofHumMI.Fadont in OyoStMt, 2011-2()12 The results from the G atatiatic for testi:Dg spatial dependence are given u folloWB: for 2011, G Index (JH(JIIES'I' POBI7'1.VE VALVES) equalled 0.19, while th& Z acore equalled 0.93 &tandard deviatioDB, with no apparent cluaterillg det«:tecl at ~ eeale. For 2012, the G i:D.clex equalled 0.42 with a Z 1100re at 6.21 &tandard dmatiOJJa, with a leaa than 1'i> likelihood that tha clu.&teril!g of high valuee could be the 1'\llult of randomchance(se&also'nlble8). Hotspots (Getiiancl Ord) There were no hotspots fbr RTO as a result at human facton! in 2011 (Figure 25). In 2012. the hotspots for RTC as a result of human factors include Oluyole, Ona Ara, Fpda., Lagelu, Ibadan North. lbadan North East, lbadan North Weet, High positive valuea for the no:rmeli..M Z valuea lbadan South East, and Ibadan So11th West LQAs for the G Statietic were o'baerved in Olu,yole, Ona (see Figure 26). ka., Egbeda, Thadan North East, lbadan North ..,...... West, Thadan South East, Ibadan South Weat, N Jhadan North ancl Lagelu LGA in 2012. The ·-· ... - .. oon.centration was only high in Oluyole LGA for the _ i year20ll. f ' e...... DISCUSSION / Over 80 percent ofthe 1.3 million people llilled by the the •· The Nigerian HNith Joumal, Vol. 13, No 4, Octr..,., December, 2013 IPage CU. I Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A. many years been neglected by global health and Furthermore, the state of the road makes it very development agendas [3]. The WHO also asserts easy for vehicles to swerve off the main road when that more than half the people killed in traffic a driver over speeds or losses control. The spill over crashes are young adults aged between 15 and effects indicated in the results is supported in 44years often the bread winners in a family [1]. reality; almost every portion of the road has an The findings support existing literature on RTC RTC record. This conforms to the FRSC RTC deaths and RTI as growing health issues (2011/2012) records, when the land marks within disproportionately affecting vulnerable groups of these hotspots are observed very closely. The road users, including the poor [2,5,7,9,10]. In the causes of RTC however, varied in other LGA but present study there was very high indication that mostly not the result of wrongful overtaking as more males sustain injuries than females as a observed in these LGAs. result of RTC. For the two periods, the number of males and adults that sustained injuries was high, Next we focus on speed limit violation, for the two when compared to the number of females and periods under study; the null hypothesis of children. Also, the study, gave a very strong randomness was rejected. Based on Moran's and indication that more male adults die as a result of Getis and Ord's statistics, we conclude that the RTC. For instance, the number of adults (blue) spatial pattern is clustered, that is, there is spill killed in 2011 and 2012 was higher than the over effects of speed limit violation within number of children killed (yellow). Again, the particular LGAs across the State. The hotspots are number of males and adults killed as a result of Oluyole, Ona Ara and Ibadan South East LGAs. RTC was high, when compared to the number of The study observed the major road network females and children killed. Male adults' death (Ibadan-Lagos expressway) within Oluyole LGA could impact negatively on the families, work force had heavy traffic flow and drivers travelling at of the nation and invariably cause a reduction in high speed. Ona Ara and Ibadan South East LGAs the GDP. Thus, increased focus on deaths and are contiguous to Oluyole LGA. As such the spill injury prevention policies is required to stem the over effect makes these two LGAS prone to RTC. trend. This road network is a busy long stretch; some parts have potholes, sharp bends and adjoining In the studies ofPlasencia and Borrell [5], Leveque roads. Looking closely at the landmarks used by et al [6], Cirera et al [7] and Labinjo et al [8], the FRSC in the RTC data, most RTC were marked differences were observed across districts observed close to T- junctions. This suggests RTC and temporal variations recorded in respect of occurs when pedestrians attempt to cross the investigations on deaths and injuries resulting expressway or vehicles attempt to join the from RTC cases. In a similar vein, this study expressway. The attitude of drivers on this observed geographic and temporal disparities on expressway could lead to losing control of the the causes and characteristics ofRTC, deaths and vehicle especially when there is an unexpected injuries. One of the several causes of RTC is need to slam the brake. dangerous driving. Considering the two periods investigated, the null hypothesis of randomness With a special focus on mechanical fault were RTC was rejected. Based on Moran's and Getis and is a result of brake failure and/or tyre burst. For Ord's statistics, we conclude that the spatial the two periods under study, the null hypothesis of pattern is clustered, that is, there is spill over randomness is rejected. Again, based on Moran's effects of dangerous driving within particular and Getis and Ord's statistics, we conclude that the LGAs across the State. The hotspots are Oyo West, spatial pattern is clustered, that is, there is spill Oyo East, Afijio, Atiba and Ibarapa East LGAs. over effects of mechanical fault within particular The study observed the road network within these LGAs across the State. The hotspots are Oluyole, LGAs is one way route from Oyo through to Ona Ara, Lagelu, Ibadan South East and Egbeda Ogbomoso town; some parts have sharp bends; LGAs. Oluyole (Ibadan-Lagos expressway) and most parts have potholes and the asphalt on some Egbeda (Ife-Ibadan expressway) LGAs have major parts are no longer regular and smooth. As a road networks with two routes and very heavy result, once a driver makes a wrong attempt to traffic, while Lagelu has a major road network overtake it leads straight to head on collision (Ibadan- Iwo expressway), which has one route. because there are no side routes or lay bays to hide However, Ibadan South East and Ona Ara LGAs for both the dangerous careless driver and the are affected because of spill over effect; RTC in innocent careful one maintaining the right lane. these areas is high because of the frequencies of The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPage IFf) Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A. RTC in the contiguous LGAs within the state. and injury prevention policies targeted on the Although, a few of the RTC cases were the result of adult males in the state. This is because of the dangerous driving, speed limit violation and significance of young male adults in the society. In human factors but the main cause in majority of particular, deaths and injuries of young male the RTC cases in these LGAs is mechanical fault. adults have a significant effect on the workforce of Increased vehicle inspection in these LGAs could the nation. This impacts negatively on economic promote safety and security on the roads. activities and can indirectly cause a devaluation of the country's GDP. Due to the socio economic Lastly, human factors, categorized as driving implication of losing young adults, further works under alcohol influence, overloading, obstruction could aim at investigating the age group of persons or bad road could cause RTC. For the two periods involved in RTC and the effect on the occupational under study, the null hypothesis of randomness is structure. Also, the cost of traffic injuries to low rejected. Based on Moran's and Getis and Ord's income and middle income countries is a global statistics, we concluded the spatial pattern is issue of concern. This has been described as being clustered, that is, there is spill over effects of higher than the total development aids received by human factors within particular LGAs across the these countries. As such, an investigation into the State. The hotspots are Oluyole, Ona Ara, Egbeda, cost of RTC will provide a good guide to measure Lagelu, Ibadan North, Ibadan North East, Ibadan the negative effect on the society. North West, Ibadan South East, and Ibadan South West LGAs. It is evident that each of these LGAs is REFERENCES contiguous to at least one other LGA in the group. 1. World Report on Road Traffic Injury Another unique feature of these hotspots is high Prevention. Summary World Health traffic volume and population density in the LGAs. Organization 2004. 2013 Mar 17; Available Again, there are major road networks across these from: URL:http://www. who.int/violence_inju LGAs, most of which are in the heart of the state. ry.. ./road .. ./world.. ./summary_en_rev. pdf Moreover, the study observed serious business 2. FRSC: Nigerian roads second worst in the activities along these routes, for instance markets, world. 2013 Sep 06; Available from: garages, schools, filling stations, banks and URL:http://www. thisdaylive.com/articles/frs hospitals are some basic features that attract c-nigerian-roads-second- worst-in- the people and can be described as trip generators world/103012/ which indirectly generate RTC. Drivers attitude, 3. Global Status Report on Road Safety. therefore need to be put under control to ensure Supporting a decade of action.[pd£].2013 Oct security oflives and properties. 20; Available from: URL:http://www.who.int/iris/bitstream/1066 In summary, RTC in Oyo State are characterized 5/78256/1/9789241564564_eng. by deaths of male adults and the same group of 4. Reynolds DJ. The cost of road accidents. J R persons get injured. This category of adults is Stat Soc SerA 1956; 119 (4): 393-408 likely to be young adults, probably bread winners 5. Plasencia A, Borrell C. Population-based of their families. Results show there is spill over study of emergency department admissions effect of causes of RTC around particular LGAs. and deaths from injuries in Barcelona, Spain: The concentration of RTC for the hotspots is high incidence, causes and severity. Eur J indicating spatial dependence. This suggests that Epidemiol1996; 12 (6): 601-10. safety and security measures must be 6. Leveque A, Humblet PC, Lagasse R. administered across the hotspots simultaneously Premature avoidable deaths by road traffic in order to achieve significant remedial effect. injuries in Belgium: trends and geographical Apparently, RTC are mostly the result of drivers' disparities. Eur J Epidemiol 2001; 17 (9): actions or inactions, thus, these attitudes can be 841-45. tailored positively to enhance a reduction in the 7. Cirera E, Plasencia A, Ferrando J, Segui frequencies of RTC across the state's roads. Gomez M. Factors associated with severity Furthermore, there is a need to improve the and hospital admission of motor-vehicle conditions of road networks in the state if RTC injury cases in a Southern European urban cases must be reduced. area. EurJEpidemiol2001; 17 (3): 201-08. 8. Labinjo M, Juillard C, Kobusingye OC, CONCLUSION Hyder AA. The burden of road traffic injuries The results should enable the orientation of deaths in Nigeria: results of a population based The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPagelf:!.i Spatio- Temporal Analysis of Road Traffic Crashes- Korter G.O, Olubusoye O.E, Salisu A.A. survey. Inj Prev 2009; 15(3): 157-62. 12. Ogbodo D, Nduoma E. FRSC: Nigerian roads 9. Chandran A, Kahn G, Sousa T, Pechansky F, second worst in the world. 2013 Mar 17; Bishai DM, Hyder AA. Impact of road traffic Available from: deaths on expected years of life lost and URL:http://www. thisdaylive.com/articles/frs reduction in life expectancy in Brazil. c-nigerian -roads-second- worst-in -the Demography 2013; 50(1): 229-36. world/103012/ 10. Ezel UO, Kipsaina CC, Ozanne-Smith. Fatal 13. Sukhai A, Jones AP, Love BS, Haynes R. road traffic injuries in Ibadan, using the Temporal variations in road traffic fatalities mortuary as a data source. Inj Prev 2013; in South Africa. Accid Anal Prev 2011; doi: 10.1136 I injuryprev-20 12-040674 43:421-28. 11. Federal Road Safety Corps, Federal Republic 14. Moran P. The interpretation of statistical of Nigeria. Road traffic crashes data. 2013 maps. J R Stat Soc Series B Stat Methodol Mar 17; Available from: from 1948; 10:243-51. URL:http://www.frsc.gov.ng/road-traffic 15. Getis A, Ord JK. The analysis of spatial crashes- data association by use of distance statistics. Geogr Anal1992; 24:189-206. The Nigerian Health Journal, Vol. 13, No 4, October- December, 2013 IPage If: II