Johannesburg, South Africa. The aims of the study in South Korea and Cameroon in the same year were were (1) to determine the incidence rates per 100,000 ...
s,c. S.,. Mdr. vol. ll, No r, pp ls] 237, Print d in Gr.ar Bndn. All tshb r6(ved
r9l
0277-9516j9r Sr.0O + 0.00 CoDyriEht
O l99l P.rFmon PEis Pl.
EPIDEMIOLOGY OF TRAUMATIC BRAIN INJURY IN JOHANNESBURG-I. METHODOLOGICAL ISSUES IN A DEVELOPINC COUNTRY CONTEXT DIGBY S-
O
BRowN and vrcroR NELL+ PO Box 392, Pr'loria South
Heal$ Psychologt Unit, Uoilersily olsoulh Afii@,
Aftie
in lhedesloPinscounrriA pr een I Aidenroloei'Ls wil h unique f,dhodololi_ solur.on) m usr b€ lound The* challenges a n* lion at'hr @ l rndeq @cr\' e) lor cosr epi' Eres, and lin ed esrch lLndrns rhar t'i*t' rhe n. rhod' dd ised in Johann$bu rg. sourh A rnq in order to devdoP o.i 'iolence I."i.1."". i r''i Eliab'l vorLhe morbldirvand monalnv ansrns fiom traumarn bnin Liiiii.d *'1.",'", "r r€ported in Pan Il oflhe Psp€r G. 289) Rsults ae irs erioloey. injury, and
Abcmcr lnc'dere rud,e'
el
challenses for whKh
no!.
nssiu
i.,ii,"r ."'*"*ai'*. ,i*: -*i ***.aole
(€), Bordr +lldemiology, nelhodologv, traunatic bEin injury' saDpling,
This two-pa.t paper enmin€s the epidemioloSv of faral and noniatal lraumalrc b,ain rnjury ITBIr in JohaDnesburs, South Africa. The aims of lhe study sere lll lo derermine Lhe incidene rates per 100.000 of boih talal and nonldtal TSI in rhe lohannesburg mashrerialdarricl in 198o, t2) ro rdenlilt lho\e age. mi ana *x sroups aL Dosr r(k ol TBl. (lr lo
olTBI, r4r ro detemine 'dorifv wrrar itre ciuse' or TBI $ere in .pecrfrc age. race lhe orox'ma(e cause.
and sex rroups, and (5) to develop bypolheses thal aeount for lhe disrribL,lron dnd cause' ol TBI ;n Johannesbuis. as tevealed bv lhe studv Pdn I reviss the merhod olosel bsues Ihar arise in this conlext of underdevelopment and neocolonialism. Pan II rcviews the rEsults of the studv on the morbidiry, morrahry and eriolog' ol T8I in JohannesburS. 'Numerou" merhodological Fsues ari5e lor epi_ demiologsti in de\elodns countries one ol thec. rhe rae segre8aLion olpublc hospirals t is unique to South Alrica. Others are common lo all developing countries (and some developed counlries), for €xamPle
,ncomplele and uD'eliable archival hlormat'on in hosp,lal rccords: rnadequsre re!earch lunding 'haL necesitat€s low @sr eoidemroloEv, a colon;al legacv that has sk€wed healrh;are structures in favour ofthe colonial rulers and their suc@ssors, so thar manv
oublic *cror hosprLali are ovetcrowded and undeF ll;and lhe veq hr8h rarer o'violence related
iraR-ed
iniury and dealh in these countries.
The ohenomenor of excessile violence in neocolonial'societid is well known t2. ll, and documented in Pad
II
of thk paper. Road 1rafrc accidenls, the
sinslr large"r contriburor ro rtaumatrc brain iijurv in all counlrrer, run dl exceprional\ hr8h rates in manv of the developing counlries of Af.ica and Asia. The Internaliond Road Tramc Fede.ation {4lrcpons that faralities per hundred million vehicl€ kilometres in rhe
.To whom correspondc.ce and leprinr requ6rs should bc lRacial segregation of South African hospilals was onciallv abolished on
I
I May
1990.
developing
'ohan'esburg,
United Slates, Great Britait and Japatr in 1985 were. resp€crively. L5. 1.8 and 2.2. ln contrasl. rhe rates h South Korea and Cameroon in rhe iame vear *ere 54, and 53i in Kenya atrd South Africa in 1984' 39
dd 18.1, r€speciively. lo the developjng coun(ries. the Dsvchosoc'al Dtoblems laced bv
monumenlal
TBI su^ivo6
are
i"l"e.tatea d""use menr"l heahh and rehabrltLadon orofe"sionals are eilher h :horr :upplv or enrrrcl) unavailable. The disabled are poorl' *ned bv lhe
healih care system, and the rnentally ill including TBI survivors-are the poorest of the poor [5,6]. Thele problems mun be addre'sed by largescaleskiils de\olulion programmes within a con(etr ol cost effective primary health care Motivating and sustarn-
ine rhrs trocess requires ac.urare dala on rhe scale of rhe.ublic health oroblem oreserl€d bv TBl, and lhe rare'ar which (his problem is gro*;tr8. sinc. survival increases as tEuma care imProves. Growth in TBI survival rales wll b€ gr€ler in these etLings lhan 'n care is alreadv highlv
countries
in which lmuma
i. s""tl af"o
lhere have been orher ltudEs ol l?-101, buL some have dealt onl) qrh children. and none hdve developed rncrdence rares b) colenng all adult cases in a known PoPulalion, as in the present srudy. The de\elopment ol accuEl. enimatcs oITBI incidene a( low co'L required lhaL prospelive.
TBI
bedside data Salhering b€ substiruted for the mo.e con\enrional reLrosoecrrve archiral methods usuallv louowed in developei counrrie'. becaur essenhalddla were often missing f.om hospilal records. A final and forlunately Purely l@al issue lhal Posed melhodologiql p.oblems is the Soulh African govehmenfs widely condemned policy of Eciallv Fgregaled heallh ca.e rh3t results in what has been termed '1he aoaflheid disease ll Il. ln rha counrrv a le8acv of
c;lonial health care has been delibe'.reD 'h.rpencd
by stalulory dilcrimination and the fragmef,tation of health care servi@s 021. so rhat rae has become an imDorLanr delerm,nant oi 5ocroecotomic nalus and of;ccess Lo heallh ca'e serviLe and lhcir quahlv t hc
2U
DGBY S. O. BRowN and
health slarus ol South Africa\ Biack* poputarion, conprising 84% of the country s total popularion, and 80% ol tbe Johannesburg populatioq, is thus The epidem,ologicdl iu^ev merhods dcvetoped tor thn $udy. and rhe i\sues it r.sohed rhus md) be ot interesl Lo eprdeniologisl, and public hedlth (prfidlisls in lhe Udled Sutes and Europe who conducl or review Dublic healr\ rescarch in lhe devetoping counlries, as well as to heallh care researcheB and
policy rnakes
i!
the developins counrries.
I, MORBIDITY: Mf,THOD Subjects
The target popularion was all residenrs of the Johannesburg magisle.ial disr.ict aged 15 or nore. The su ey populaiion was inparierts at carchment hospitak meeting case admission criteda for nedic.al eligjbiliiy, incidence, ard ptac€ of residenc€. The cdch-
nent hospitats were all hospilah adhilting trauma cases.
i!
Johannesburg
Medical eligibility ffirerid denred, wirh modrficrtions, from ihe Head and Spinal Co.d Injury (HSCI) survey commissioned by the U.S. Nationat Institutes of HealLn .lll. and specrhcdllr rhe me,hodolosical chaprcr ol rha' .Lrvey ll4l Case rncturion reau;Ed an inpa.rcnr ddmi$ion as o direcr re\ulr oi a Tat, defined as cerebral contusion or laceration. with or wilbour the loss ofconsciousnss, or iiauma related unconsciousness or amresia of whatever duralior. These included codes ll4l are defined by the.arure or injury codes (N-cods) or rh€ Inlemational Classification of Diseases n5l, namely N850 ro N854. aDd nental disorders frequently aslociated with kauma (N293.0. N293.1, N294.0, N310.0-N310.9). The.case findirs codes' raise the suspicion ot TBI: these are slull lraflur.s r^800-N804r. cranial nerye injunes fN8?0-N871. N950 Nasj): uarmafic complicarions
(N958.1-\9r8-lr. and nenous sJftm or
endocrine
di*ases (N345, N347, N348, N253.9). Individuals wirh case-findrns coder were rncluded ii Mrhin five days there had been an episode of unconsciousne$. seizures, headaches, vomiring, or cEfebrospinal nuid Only fi6t adnissios rere eligible: discbarged cas€s lor la1e complicadons were excluded. Interhospital transfers with no inlenuirg discharge
readinitted
While and l2% ofBlack admhsions ro Johann€sburg hospilals were nonresidents. though cdse kakage in the revese direclion is likely to have b€en les_
Incidere .ate is defined 6 the numbe. of rcw
qses ofa disas pe. 100,000 ofa known population, occurring in acalenda. year. Here rhe popularion was Joharnesburg .esidents in 1986. Altboush prevalence is a common epidemiolosiel mea'ure. rl was nor used in rhis nudy. Pr.valene is defined as incidence x the average duraljon of dis€ase, and in rhe conlexr of TBI, is virtua y impossible lo dete.mine the 'duration'! of dis€ase.' since this nec€sitates lpe.ifying how the mulrifacroriat mturc of rccovery after TBI is to be defined, and ar what poinr rn Lrme rhe parient recorered a(ordinc lo lhe chosen cnre a ll6l Unril Fco\en afier TBI can be
adequately defined, TBI preval;ce fisures will be Incidence cases were those admitled on the surrey
day o. precedins day (se. 'Samplins p.oc€dure, below)i cases admitled p.ior to rhis were defired as plevaleht and excluded fron the count_ Sin@ case identifietion ward .ounds took place between 6a.n. and 8 a.m., elapsed tine from trauma ro case identi6, cation !aned belween I hr {the mrn'mum for proe$. jne a case rhrough ambuJance dnd emereency room to the ward), and 32 hr. De te rnin ing po pulat
ion
size
A reli.ble estimare ol rhe race. *x and agr d6rnburion of Johannsburss popularion s Fquired to .alculaLe lhe incidence rare wnil. rhe 1985 cehus
fisures fo. Johannesburs\ !71 Asians, Coloreds and Whites have not been chall€nsed, rhecensuscounr of 809,605 Africans has b€en rejeled as too low by borh govemmenl and indep€ndent sources.
Tbe population of Soweb, a resion ot suburbs within Johannesbu.s in which (lesally) only Africans rnay live, has b€€n estimated from 1,069,08? [t8] to 1,470.000 9l- Of the available esrimates. the best appean to be 1,350.000 based on effiuent measured lrom Soweto by Planacr. a
Lown planning and engrn-
eering group with no vested inlerest
in over, or
underesrimarinS rhese figures. To rhis ngure musr be added rhe number oi Africans living in Whrte' Johannesburg: A variety of estimate! cooverge on
I50.000. grvinB
a toldl ol
t,500,000 Atrican\ ,1
hospi{al in which they were
Johannesburg in 1986. Incidence rar€s lor Af.icaDs were based on rhi( fiCure. and rhe age and se\ dislriburion determ,ned b! rhe 1985 ensus aDDtied to
Onl) nerron: reidenr in johdnne'bu'e lor ar kajr
The 1985 census fisures were adjusred io. 1986 by lhe estimaled annual giowlh .ates of 2.64% for
were counted
in rhe
Geagrcphic eligibilio J
vroon NrL
mon'h. pnor ro rniur! rere adm ,ed ro lh. incrden(
count: a separalc counr was kept ofnonresidenls. The incidence figu.es are an underestimale, b€cause no enimdrc of brd,n rnjured e(idenls Jdmilred ro hospr. lah oulside Johannesbu.g is available. Some 26% of
'As olbg dse.rcheG hale noied. rhe €minolosy tor Sourh Airica:ra.csroup\hd ril"hetd lheu,aseh(,( r. A nc"r' ro ddrgnare Bdnrl .pa\rng peopt;. r'nd
'B"ct ro' Alncrn.
Colored tm,\N Ecer and Asi4n
Asians, 2.6% for Asians. 2.6% 1.85% for
whils
lor Coto.eds, and [20]. Census undercounr estimates
were not applied
sine they vere
calcutaled ior
rhe entire counoy 7l. and rbe undercounr varies considerably from resioD ro region it8l.
Injur!
te"e tn y
c btsifcat ioa Major difrculties atlehd deErminarion of severity of TBI. usually denn€d by duradon or deplh of un,
conlciousness. ordumlron olposr-traumd' c amnesia. An initially mild brain iniury may hcome selere as a
Epideni.losy of lraunalic hrain injury in Johan.csburs I
Lbl. L Ho.p
resulr of secondary pathologic evenrs. such as septi_ cqnja [21]. 'Ptimary a.d s€condary' severity were nor drsutreuiJed sinc€ end-:laresere r\ is lhevadable ol inrere-sl A turther conloundtnc itcro, arises fiom rhc Dracr;ce in some hospirah ol nedvily sedaling TBI ;ares d( a prophJlacric mea(Lre lor a l-5 day penod piolorging ucon$iousness artifically. Howeve., this would rarely be apPlied 10 light or moderale cases. so
it
st. ! st cld$ iid suft) dr)s
Brds
Visir Sury.Y Rae l23l clcl6 daYs ddrsna.ion 39
does not unduly skew the selenty distribut'on
Glassow comi Sqle IGCS: [)21] oi unconsciouenes qrh s scorc lhdr varies bel;een 3 and 15. Although inlernationallv accepred 12ll, lhere b no unanrm rl on which CaS
Derl, Thc
t2
210
inderes .lenLh
{o,e. mark the bounoar'.s
b€rween
r\erirl
le\ek
For a $vere i.jury, Kalisky er ar [24]su88est a score of 8 or lesi winosram er ar [25] Propose 7 or less.
In
@nsullation wilh the hospilah con@rned. lhe dema.caiion adopted lor thc present sludy was a cCS scor€ ol6 or less to indicale severe injury. 7 12 for moderale and 13 15 for ligh! injuries
Drrar',,. Unconsciousness' was defined as a GCS score oa 12 or less. Three severity gtades were sp€cified:
of l-59 min was defined as a mild 24hr as moderate, and 24h! or more
unconsciousness
bnin iniury.
I
Post-truunoti. ana.sid
(P"l)
This is defined
as
rhe penod b€lwtrD inJUD and the re'um oiconlinuous
remor r, ds dr\rrncr fton 'iolaled iJands ol recall is usually some lour limes tonger lhan unconsciousness dumdon 1261. nrc palienr's perceplron ollhc dLralion of PTA lypicaily increases betwcen lhe first return of consciouness and lhe end of the spomdeous reco\€rv Denod .ome J to o monLh" po(r-rra"ma [)7] 5ince one period of mnesir may be nes'ed qLhin a ronSer period. ln consequence.
Lhe
Pdrienrr RsPon)e
ro h.
quesrion. When did )ou {al'e
up or re8a'n \.onscr^r h_ ness?" is inaccurate id the hours a.d days Posrimuma Caution therelore rcquires that the distincljon beiween
lisht and moderate gBdes ol severity be
recognized
,A: Afri€nsiAs: AsirnsiC: Colortds: w: Whnes 'vniL ro pr\,e hdp tk roor Dae con\u .n'l) r:rh publi! dd rc rhe rokl ; i.'cre. o'
cases seen
in a year. Accordingly, one of the five for I wek, and
public hospitals was visited daily
there we.e nine such hospital visit cycles of 5 weeks The enunerator worked 6-day weeks, with a 4 dav break, Friday to Monday. at lhe end of each cvcle; weeks 2 4 of each cycle, the day of followed a
i!
Tuesday wednesday Thusday squene- From the fiflh cycle, each private hospital was linked wilh a round at a public hospital, so that all eight hosPitah could be iDcluded. However, becau* of scheduling e(o6, visit cycles per hospital varied b€tween 7 and 9 Oable l). Durine i tial 6-day cycles at lwo private
hospirak (Garden Cir) and Vrlparl). no rncident rni cases were locaLed. and lhese were nor vrired
'The seouerce in shicb hosDirals were visrred in eacr cvcli was.andomrzed. Ar each hoso.lal, lhe ward visit routine developed durjns the pilot study {as tollowed. lncidenL cases rhar died after admi*ion bur Drior to our ward round. would have been missed
fron the
count- Such cases would not have been
missed by experimenlal designs such as lhe HSCI suryey Ll4l. No rstimare of this eDor wa! made. bur
posrtraumalic amnesia nore often thar on unconsciolsress duration.
rr is unlikely rhar
All cases were rated by annesia-unconsciousness duration and GCS score as mild, moderate o. levere
dge,
as based on
injuries- Unconscious Flien$ were rated as moderate or severe using GCS rcole. Consc'ous Patients wcre rated as light or moderale usin8 esdmaled PTA or dnration of uncon$iousness data. Procedute
The main sludy was pr€ceded bi' a 9-sek pilot srudy, fiom Varch J ro May 2 lo8o. dunn8 which
archi\dl (amplhc $a' Jnderra\en ard rej(led ds tnadequare. ind a qard (alcl'ment toJr,ne detrrd lor edlh ho(pi{al 'uch lhal 10 nes TBI cdse. $ould be Sonplihg procedutc The 4l-week main study ran from 5 May 1986 to
27 Februa.y 1987. Thc liniled furding available allowed employmenl ofone field sorker' A nelhod had 10 be devised thar would allos sampling ol all
ol ric\ olrhe numbcrolne\\ (incrdcnt)TBI
cighr hospiral and rhe detclopmen or esrimale: knnwn
^m
il si8nrfiqnrl)
affected Lhe 6ndings
The inBke protocol gave identifyitg data {name, fr, race. Ianguage. prace and duralron ol te\i denej, admision data rdale. admlting unrr. diagnos; 1n.i rherhe' .r nor los, ol connioutnesr had been noted on admission), and iniur! inlonnation (date a.d lime oiinjury, CCS and !h€ time it was taken, period of annesiaronsciousness, and the ICD-9 codes ior narure and causes of inju.y). GlEcaw Cona Scale. GCS scores were assisned on 6rst conlacl with a case. Siie GCS may improve rapidly. the elapsed lime from occur.en@ of lrauna ro GCS scoring must be tak into account For incidenl cases, 0.9% were assessed b€lween I 3 hr, 2.3% berween 3 6hr,30.8%6 12hr,355% 12 18hr and 30.6% be$€en l8 32hr. Since elapsd lime lo CCS rating was 12 32h. lor 66.1% of the sample and 6 hr or more for 96.9%. mosr cases were rared after initial stadlizalion. Data Ahat$is Deriving incidence rale requ;es the annual number cases of TBL FoT each ase-, sex-, and race-
of
Drc3Y S. O. BrowN and
246
Nmber of cases was divided by the number of'he smplins days on which those cases had b€ln found. and lhe quorienr multiplied by 365.25. Since ahe number of sanpling days at each hospiral ditrered, annualized fiCrrres were 6!s1 calculared fo. each hospital sparalely. then summed. category,
Ac.utac! of anaualized esrinotes Followins HSCI Il4l, the rclative slandard eror (RSE) was uled to detemide the accuracy of the brlalizen eslimates ofTBL RSE i! used in prefer€n€ to 95% corfiderce inlervals. The RSE ofan annualized estimate is obtained by dividing rhe standard edor ofthe estimale by the estinare it.elf. and multiplying by 100 to obtain RSE as a percentase. This allows the relaiive accqracy of aDnual esrimaies of differenl maSnirudes, but the same srandard enor,
V'sro* NELI
pathologisl had specified brai, as cause of death'on the autopsy report. Fo.o(ample, 'njury il'cardiac
an€st (stab heart'were given as the caus€ ofdeath, a concurreDl finding such as jnlracerebral conruson wonld lead to case inclusion. However. when c€rebral pathology irdi.aled an equivo€l eliology (erebral i.farct'), to which ar included N{ode could ror b€ applied withoul coEoboraling evidene of a lraumatic event, the case was excluded. Also omiued wer€ traumatic injuries described as'old,' rhat is predatin8 death by two nonths or more. Since autopsy evid€nce is not equivaleni to clinical evidence, this proc€dure admifted ro the count cascs that ma, nol have met Lhe medicll eligibilir) cnrena wilh equivalenr bul nonfa(al TBI Neve heless. gi!en the frequently unequivocal and massiee brain damage recorded at autopsy, the overcount is unlikely to have
ao t'e compared. whereas the 95% confdence inrerval doesnot. The smaller the RSE, the more a@urate rhe
annualizd 6timats. Sinc€ va.jance is emple sp€cific, RSE\ a.e nol comparable across studies. The RSE also provides a partial ind€x of the magnitude of the eror associated with the computed incidence rate per 100,0{}0t the r€mainder of the error in incidence .ate is due to inac.uracies in estimaring population cell Calculation of the RSE must lake hospilal race desi8nation into ac.ount. Under nomal circunBtancd the number of suwey days (X) would have b€en 230, the total numb€r of sanpling days ove! which the study ran. However, a RSEcalculaled foragiven ser. age and race category usiag total suney days in the denominator would b€ spuriously low, since n would include vasit days to the bospitals designated for othe. ra@s, on which the probability olfinding cas€s belonsing to thal mce would be systematically reduced. Inded, our data indicated thal case leakage acros hospital .ace designations was very small. AccordiDgly, since the sharpest race distirction was between Whites and Blacks, N was se! lo 88 for Whites (rhe numtler of suney days at While d6ignated hospitals),
and
to
142 for Blacks. To estimale all-rae TBI
accuracy, th€ lotal number ol survey days was used. 2. VORTALITY: METHOD
AII knowr or susp€cted
cases
ofunnaruml dearh,
which includesTBlderh(. are altoFied: rbs euminalion routin€ly includes a coronal brain seclion. From mo.luary data of all deaths belween I February and 30 June 1986. separale counls were obtained ol Johannesburs residents. adllt.esidents (a8ed 15 or more). and aduh residents wirh bra'n idjury. Sin.e rhe hosprral number oi all hospiral dearhs is recorded in thenonua.y file, a crossche.k againsr th€ incid€nl nonfatal cases was runj paden! names were also c.osschecked. No duplicares were found. so thar co..ect TBI incidence colld be derived by sunming fatal and nonfalal counls.
TBI eligibilit!
All cases wilh unequivocal auloply evidence of traumalic brain danage, 1o *hich one o. more of fte HSCI 'included N{odes could be appl;ed. were admilled to the court, whether or no! rhe examining
Estimates
of the annual nunb€r of TBI related
deaths were obtained by dividins the numb€r ofcares by sufley days and mulliplying lhe quorient by 365.25. Sin@ morluaries are nons€grcgared, the annualizrion and reliability issues reviewed for nonfaial TBI were irapplicable. As no1ed, RSES for fatal and nonfatal
data cannol be compaEd, since vadances ditrered. 3. Ertor_ocY
For both the morbidity and horrality srudies tCD-9 codes ll\l we.e used ro define narure ol injury (N-codes) and €xternal quse of injury (E-codet. Data quality checks in the fourth monlh ol lhe
morbidity study rev€aled codins inaccuracjes. and con*qxently E and N codes gathered during the first four months we.€ discarded lrom the dala!€t, and reimining of field workers impl€mented until 90% ac.uracy was achieved. Acknodledgenents We
Sraleful !o rhe Insriture for
^rcthe r{una. ScienB R6.arch R6eaEh Developnenl ol Council for fnndirg; to the Tracvaal Provincial Hospirll Adminisrndon and lhc privat€ hospital nanagementsj 1bc sraff and palienrs ol lhe eighl hospn.h and 10 rh. sraf of rhe Johannesbure Inqucsl Courr. W€ a.€ sralelnl lor 1bc inlomed inputs made by Proja! Adviso.y Conmilte mmbers under lhe chai.peNonsbip ol Dr K. F. Mauer,
Gp€ially
ro D6 D€Et Yacn. Bemard Fourie, Patrick Hs*1, Percr MiU.r. and Roelf Prinsloo. Mr Andrew Chiloane and Ms Brcnda Radebe gathered rhe da1!. Th. comnenls of rwo a.onynous Eviewe6 are grdteaully acknoyledged. The viws exp.essed in this pap€r a.e tho* oflhe autho*, dd should nol nqesarily b€ atl'ibuld ro th. funding or p.mf sron-emnnne hso!utrons.
REFERENCES
L
Sa"dcts D. The Strussle lu t985. '1?a./rr.
M
a@illan. London,
2. Auslin R. The colonia! nod€I, subcuhural thsry, and inters.oup viDlc..e. .L
I 4
C,irkall6tice
ll,9l
104,
1981.
Aulrn R.Pro8re$rosrro.aci,l.qJahryandreduclron criminal vialanc., l. Oininal J6tice 15.
ol bh.k
437 459,1941.
lnremaroml Ro.d Fcd.rd
/9r.t-l982 Aulhor,
iot
wo4d Road StahlL,.
C€neva, 1988.
EpidemioloCy
1. world Hea
counnies
rh
/
ol tdumalic brain iojury in Joha.n€sburs-l
Orsan$mn. u tot H.ahh in DNloPnA Oit;col App,aisal of Rdeanh Fi\dinet'
Teinical Reporl No. 698. Allhor, G.t.va, 1984. 6. World H€lth O.ganiaion - Poblerc of Mental H.alth in the Aiicah Reqion. Anthor, Br@v le, 1986 ?. Austen M.. McLeUan J., Yach D and Knobel G J Falal head injuries in Cap€ Tow SA Med. J 7\ 710 172, t9at. 8. DeVilliers 1. C,. Jambs M,, Parry C- D. H. a Botha .t. L. A rclrospedve study of head itjured childnn admilled 10 two hospilals
9
il
CapeTow.. S.A. Med J-
{05, 1984, Ro*-hn€s A. P.. Le Roux C. J.
66, 801
C
In0u.ne ol
road speed re,rncuons on rhe:ncidereofh.ad rniuries 254a 2551,1914, S.A. Med. J10. Van Delten.,. R. Annual slaiislics. D€Danm.nl of sursery, Wenlworlh Hospilal, Nalal- Unp$lished rcporl.
$'
ll. De Bee. C. T,e
Soulh A-fiic@ Db?ue: APa h.id
Health Johannesburg, 1984.
12. Botha J.
L., Biadshas D., Gonin R-
18.
Kok P C.
R.po No r986.
Uoo"tlikp
1985,
Or 8) Oe (rovetnhenl Prirrer. Prelotia.
ReDatkinas Hene.spteidina
tnskm /u?rt6rr
in Die
RSA:
(PoPulauon EdNtri'
Afne Possible lonS_ Sciences Research Council,
bulion in the Republic of South
t.m inpli@tions). Human
19.
ol cooslilulional Aflairs and Pla.ning. Dar 2a M^y, t941.
Minisler Bdiness
20, Kusher H. G. V. The 1985 @asus. Epid.niol. CMmts \4,2 44, 1981. 21. Pang D. Parholosic 6rrelat6 ol ndrob€havioural
synoromd iollorrng clo*d head inju.y ln lt dd Intw, Rehonihlahor: arildrca tl Adalet.e'ts tEdited
by Ylv'sak- M.). Taylo. ard Francis, t985. 22, Teddale G. and Jen.elt B. Asse$n€nt
inn:ired onsciousnes.
tl Ml',u. c D lrrolm
l,rc.r ii.8li4.
London,
ol cona and 1974.
bonl rocon pae o'lrcoEe rouoqing 'nkmauonaldata sverc lad irjury io ditr€renr
entres. Stzr Med 5, I0l-112, 2a.
1986.
K.l6ky 2.. Moman D P.. Mete'" C A dnd Van Liukn A Medical p'oblems en.ounreted duang
sl
Rhabihation of pari.nll wnh h..a injury' Areh Phrt tlerl. R.habil. 66 25 29, 1945. 25. WrnoE€n H. w., Knish'. R. V. and Ba$den H Neur;osnal deficits lolow:ns head injuMn ch,ldFn
t980.
J..lih. Newopslchol, 6, 269 2a9, 1984. 26, Gulhkelch A. N. Posl Iraunatic amnesia, posl co.cus sional symproms, ald acident rerrcsis Acla neutu.hir Snppl. I, 120 123, l9?9-
s2. 1980. 14 ande*dn D. W.. Kalsbeel( W. D. and Hartrell T- D. injury sun€v Dsie, fhe nalronalhad and 'ginalcord add delhodolos}. J, d@r,ers. Suppl, 53, Sil-SI8, 15.
C.nlral Statisri€l S.ttie. Populatioi Cenes
and Yach D
ol heallb n*ds and $di6 in Soulh Aftie. S,.. ,t i /e4 26 851 454. 1988. Add.so! D. w. and Mclaunn R. L. The.ational hqd and spinal erd injury surey. L ,ve!/osrg. S!ppl,, 53, The disfibution
ll.
17.
Nclc
dd Hakh S.tui.2s.Sollhem Alrican Rearch
281
world Health Or83n;sation. Muudl ofthe Inte atiual Stotist'Jat ctwifcation of Dne6a, tniu.ies ad Cauvs o/ Dsurr, Ninth rcvnio4 Vols l-2 Author, O.neva,
l97l.
16. Brown D. S. O. and Nell V, RaoDuy
foh
DifBe
Taunatic Bak lnjurJ in Johannesbutq: A Concu qt Prcstectiae Study, H.aftn Psychology Unit Tcchnical Repon 90/4. Urivdsily of South Africa, Preroia, 1990
27
,ansen
c. P Ih. pncho(dial con*quences ol brdin A lollo$-up s(udy ol tndivid!al: and rheir
rrauma:
ranilies. Maneis rheis, Un,v.brly of Sourh
4fi4.
28. Eneelha.t H. (Ed.) Itogital and Nuaing v.afiaok af
Solrrn, ,{y'i.a. Tlomso.,
Johannesburs, 1986
