Progress report on applied modelling in the OCP

J I|ORLD EEALTE OBCAIIIZATION aPage I ORGAT{ISATIOI| UOflDIALB DE LA .SANTE ONCIIOCERCIASIS CONTROL PROCRAMME PROCR.AIIIIB DE LUTTE CONTRE LIOI{CHOCBRCOSE IN EN I{EST AFRICA AFRIQUE DE L'OUEST EXPERTS Seventh Banako, ADVISORY COUMITTEE Eee I lng 16-20 June 1986 0cP/EAC7.5 ORIGINAL : ENGLISH eady corrpleted and varlous peraueters ln able knowledge from estLuates fron the epldeulologlcal a trlal and errord,for thc predlcted de'rad thle geve EoEs rcpronebly rG11 PROGRESS R,EPORT ON APPLIED }IODELLING IN THE OCP BACTGBOUND 1- The OCP enbarked on epldeulonetrlc nodelllng at the end of 1983, uhen a slupIe force of lnfectlon nodel was developed aa a tool for the reorlentatlon of the analysls of the eptdeulologlcal data. The results 'rere presented to the EAC rleetlng of 1984, whlch expreesed conslderable'nterest and recomnended more extenslve work on modelllng ln the OCp. .ollowlng thls recoutEendatlon a number of rueeElngs we,re held between OCp sEaff frorn varlous unlts and speclallsts 1n the fteld,of epldenloner,rlc modeIIlng fron the unlversltles of Rotterdau and Tueblngen. yosttnporEant was the rueetlng hetd 1n Earch 1985 1n 0uagadougou, where aproposal uas fornulated for further'work on applled nodelllng and lts utlltzatlon tn the analyels and planntng ln the ocp. 2. To aasure that uodelltng would reualn applled and relevant to the Prograomer lt uaa proposed to contlnue ln phases and 1n synchrotl lzatlonwlth the Progreas ln the analysls of Ehe various evaluatlon data. Thlstrould enable croBs-fertlllzatlon between nodelllng and analysls,thorough tesElng of uodels and utlllzatlon of modelling ln dataanalysis. Three uaJot phases uere propoeed:I- the developnent of a hoet-paraslte uodel.II- the development of a yector-parealte lodel.III- the coublnatLon of both nodels lnto a tranEElsslon uode1. '-'3--'At--that-tlEf{lte-{:ere*o?o€nt of a hoEt-paraslte model had alreadytarted uhlle the developuent of the vector-paraslEe uodel was to begln .rs soon as the detalled analysls of the entomologlcal data would takeoff . 0nly after thorough teetlng and utllLzatlon of these two submodcrswere they to be conblned lnto e full transulsslon model. 4. In f985 Ehls proposal waa foruarded to the EAC whlch supported tt,recoDEended further work on applled uodel1lng and requesEed thePrograEEe to report back at 1ts sesslon ln 1985. The presenE report ts1n responee to thts requeet. HOST.PARASITE I{ODEL. 5. A flrst host-paraslte life hlstory nodel was a runnlng on the OCP uicrocoEputerB ln uay t9gO. Ththe uodel sere tnttlally quantlfted ualng the evathe llteraturer rGeults froo OCp evaluatlons andprotrauue gtaff. The uodel was then tested agalns resulte for tso vllIagee, and further fl.tted uelnprocedure. The flrst flt wae already reuarkably gtrends 1n prevalencie end average rlcrofllarlal I confl.dencc thrt lity ol Chc oodclled phrnorenr ar underatood. - 1r e 11 gu t I oo T \ 6. However, for tuo factors there uas a g,reat dlscrepancy between che predlcted and observed results. The flrst concerned Ehe ratto beEween calctfled and non-calclfted dead eorEs found 1n nodulectomles durlng the control perlod. Further sl.nulatlone lndlceted that the proportlon of dead uorua, whlch do not calclfy but whlch dlssolve wlthln a shortperlod after death, 1s far htgher than was hltherto belleved, and tlrls reallzatlon haa reaulted 1n a nodtflcatlon of the analytlcal approach t,o the nodulectoEy data. 7. The second dlscrepancy concerned the varlatlon 1n microftlarlal loads betueen tndlviduals: the obsdrved vartatlon wa6 'trany tlnes great,er thauhad been predlcted, and a qulte extrelle varlabltlty fector had to be lntroduced lnto the nodel to correct for thls. Thle extreEe vartatlon ln the lntenslty of lnfectlon beteeen lndlvlduals 1s presently recetvtng auch uore at,tentlon because 1t Eey prove to be e signlflcant factor forthe flnal epldemlologlcal trends and posslbly also for future' survelllance and control 6trategles. 8. Once the procedure of fltElng and further testtng had been satlsfactory coupleted, the post-control epldemlologlcal trends,predlcted by thls Eode1, were used a6 reference trende for the analysls of the epldeulologlcal evaluatlon data; an appllcatlon of modelllng whlch has been very useful lndeed. 9. Closely related has been the use of model11ng for the predlctlon of the future epldeulologlcal trends r,6 questlon whlch haE partlcular slgntflcance agalnst the background of the Iong terrl strategy and PreParatlon for devolutlon. Vlrtually all trends observed todste are conslstent wlth the results frou slmulattons wlth Ehe fltted parameter values and the assuEptlon of cotaplete lnterruptlon of transnlsstondurlng vector control; sluulatlons whlch predlct that onchocerctaslslnfectlons sl11 have vlrtually dtsappeared after a perlod of I5 years ofinterruptlon of transnlesion. IO. However, conslstency between data and hypothesls ls no proof of Llrehypothesls, and conslderable uork has been done to tnvestlgace thequestlon of Possible resLdual transmtsslon ustrrg sensltlvity analysis.Ir uas concluded that, 1n prevlously hyp.r endeml.c areas, llght transmlsslon could remaln undetected for many years of vector control aud posstbly nalntaln a locallzed reservotr of inIectton whlch nould necesslEate che cont, lnuatlon of vecEor corrErol or tts replacemenf by.rrt .rlcernative lnterventlon. Very pEellmlnary results of thts work were l)resented to Iast years EAC, and more complete concluslons uere glven E() che JPC tn december I985. Recent epldemlologlcal results, noEably thoscfor Ehe village of Pendie, tndtcate that thls quesElon ts not purely t,heorettcal and thaE the questlon of lIght transmtsslon and Its consequences deserves spectal atEentlon durlng the next few years. l,I. OEher planned utlllzatlon of Ehe preserrE host-paraslte model w111['nvolve;t more systematlc sensttlvlty analysls of the most lmportant Parameters as well as further ftne tunlng of the model wlth lncreaslngduratlon of control. Planned extenslons of the model 1n the near future are the tncluslon of bllndness lncldence and the aspect of natlng ofO.volvulus shtle the recent progress 1n paraslEologlcal reBearch mlghtpossTEiJ enable the slnulatlon of the age-specttlc fert.lltEy of the adult male uorDs. Following requests from 0CT, tt w111 aleo be tnvestlgated hoc the host-paraslte model can be modlfled to be of use 1n the plannlng of Iarge-scaIe chemotherapy. Ilowever, 1t ehould be real Lzed thst t,hts partlcular sub-uode1, shlch does not descrlbe the fu11 Erensulsslon cyc1e, ls only relevant for the uee of chenotherapy 1n secondary preventlon but not for lts use ln transolsslon control. I';rgt 3 VECTOR PANASITE MODEL L2. A flrst outllne of 8 slrnple vector-Parasite model was deslgned durlng the thtrd treettng on applled modelllng whlch Has held 1n Ouagadougou tn narch 1985. Though thls outllne uas sElll vague' 1t .oni.tned already the lnportent declslon not to attenPt the modelltng of Ehe conplete pod;latlon dynamlcs of .S.DaunoEuT? brt t,o llmtt the model to those aspecti of the vecqor' shlch are of dlrect relevance to transElsslon..d whlch afe reflected 1n the entolnologlcal evaltratlon data. Spectffcaily excluded was the modelllng of blackfly breedlng whlch was consldered do be an unreallsttc enterprlse at thls stage,. ll 13. A fourth r"latng on applled modelllng uas hel'd in decemb,er I985 at the Universlty ff Rotterdam to dlscuss detatls of the ongolng collaboratton.'tn Partlcular VCU was strongly rePresenEed 1n thts neetlng whlch had aa the maln tten on the agenda the development of the vector-peraslte nodel. After a detatled revlew of Ehe varlous relevant aspects of the vectorlal role ln Eransmlsston, a framework was made for model whtch would be suffictently reallsttc and st111 slmple enough to be valuable ln the analysls of Ehe entomologlcal evaluatlon data. L4. The nodel w111 descrlbe the lnteractlon of two Processes: the ltfehlstory of the vector and the lntake, developoenE, and release of the paraslte. Slnce breedlng w111 no't be model1ed, the llfehlstory of vectors st.arts at Ehe nulllparous stage at Ehe rloment of Ehe flrst bloodmeal and ends tlhen the fly dies, wheEher naturally or as a result of excesslve lntake of mLcrofllarlae. The nodel w111 a11ow for the combtnatlon of vector populatlons, orlginaEtng from dlfferent breedlng sltes, to-enable the slmulatlon of relnvaslon. Varlous factors affectlng paraslte lntake and developnent st11 be lncluded, wlth as t,h. most lmportant factora: Ean-vector contect, nlcrofllarlae lntake aa e functlon of the lntenslty of lnfectlon 1n nan, Ehe ProPortlon of lngested mlcrofllarlae which pass the gut wa11 and develop lnto lnfecElve larvae and the survlval of the vector. The model rr111 produce as output vartous rates on parlty, lnfectlon and lnfectlvLty uhlch u1.11 be calculated for the lnoDents Ehe bloodrneals are taken to enable Ehelr cornparlson ul th the observed results of the entomologlcal evaluatlon. 15. Several raaJor def lctencles 1n our current quantlt,ative understandlng of the vectortal slde of the trensrulsslon cycle became apparenE durlng the meet{ng. Posstbllltles of research were dlscussed and a Prtorlty tlst was made of the most lmportant questlons. Thls list has beerr foLlowed up by VCU and several studles are already underway for sone of the questions, notably those on the relatlonshlp between lntensl ry of lnfectlon of the host and lnfectlvlty rates ln the vectors. It ul11 be especlatly lmportant to quanttfy thts relatlonshlp 1n llghtIy infected hosts and hosts wlth only old Infectlons. 15. Followlng the ueetint and the subsequent recrultrnent of a Part-tlme research asslstant/prograEDer for the proJect, a sEudy has been made of the approprlateness of dlfferent modelltng technlques for the development and programutng of the vector-paraslte uodel. Two detalled proposals, one for a systen dynamlcs type of approach and a proposal based on Dacro elmulation, uere dtscugsed 1n aprll 1985 between the collaborators from Rotterdan and the OCP etatistlclan. It rras agreed that only the uecro stmulatlon approach would neet the requlrements of the decenber Eeetlng and thls propoaal sae therefore chosen ln sPlte of the fact that uodel developuent and prograuotng wt11 be conslderably '!ore conPllcated Ylth thle approach. 17. A tlroe table was adopted for the work on the vector-ParasLte tnacroslrnulatlon model: a f lrst test nodel and computer ProBramme wJ I I be complc't.cd hefort'the end of arrgrrst 1986, and w111 he tested ln sr'1ltt'mhrlr t I i i ! and october 1n the!analyele of the entonologlcal data lrr Based on the concluslons of thls testlnS' the model wl 11 "pgiia.a ln Rotterdam drrrlng the last two months of 1,986. ,lolef sl11 then belpreeented to the next general meetlng shlch cl11 be held ln January 1987 ln OuagadouSou' f'age 4 Oua gadougou . be modt fl ed and The result lng on nodel1lng I '\ \ 't \ ILI I irtrtctr descrlbes the fu11 cycle of transmlsslon ' can be di.ctlon of'future rlsks and dynamlcs of recrudescence Lnfectlon and dlsease. Such predlctlons mtght be qulte L7. 0n1Y a modeI, useful tn the Pre of onchocerclasls TRANSI{ISSION UODE valuable for a rea!.1stic plannlng of devolutlon and of alternattve I,rtervepElon str8tFgies uslng chemotherapy. It was Prevlously agreed itrrt the developcefnt of a f u11 transnlsslon nodel, uhlch would be the ,o"a complicated tb develop and Eo test, should awalt the experlences "itf, the host-parastte and the vector-paraslte submodels. However, thefourth meetlng declded that, glven the need for and the complexlty of fulI transmisslon nodelr lt would be tlnely to start exploratory work the near future' Ig. The neetlng revfewed the maln transmlsslon facEors of onchocerclasis and ranked then !ccordlng to thelr quantit,atlve lmportance. It was agreed L.r start this year already on a f lrst, very prellmlnar,v, tialr.snission model ehlch would consld-er only the most lmportant f actors. Tlre objectlve nould 5e to develop .'rtd"1 for exploratory and didactlc DUrr)oses whtch would also faclIltate dtscusslons durtng the nexE general ,oeettr,g on modeIIln5:,. It should not yet be seen as a serious atEempt to arrtve ai a practlcally useful transmLsslon uodel. CONCLUS ION I I9. The progress made ln applted nodel1lng has been accordlng to plan: the host-paraslte nodel has been conpleted and ls belng used extenslvely ln [5e analysts of the eplderntologlcal data, the development and pro.qramiling of a v€r1tor-ParasLte nodel ts underway and preparaClons have i".*:lade for the dcselopnent of a Prellolnary verslon of the Er.irsr.l s:ion itodel' :I. Thr, hr)st-raraslte:uodel has proved very valuable ln the analysis of 3tre e1.l .'emiol.)gical lata and lt 1s expected that the vecEor paraslte -,rdti - itI hc, .tf eorrally practl.cal value for the detalled analysls of ;11 ., r.,..!:_. ,r1ol.>1ii.:.rl i 'ta. It ls much ntore df f f lcuit to give a prognosts of hu.; *i.rr.lI th.J uork an the transmlssLon oodel wtII be because of the F:-..lcLrr rluccEtaintl.ss tnvolved and the foreseeable dffflcrrlttes ln tr,--..:r,r: :his'roCel. Scwever, tE 1S alnays ver), useful for plannlng :,r:r-!{rs€. to h:u,'rrE.'+rtalntles clearly ldentlfled and thetr rclrLtve in,;.irt.?..dc pr.i!c-rly declared; a Procedure whtch 1s an lntegral ParL of qro'.lri c,3veloif4re;rL. :'hts ln ltself ts already a sufflclent justlftcat,Lon i,r= lle preli:i'rart vork and especlally for the Chorough, jolnt iii;ru*s-'urr oI tire t -.rnsmlsslo:l ruodel by Staff members from all OCP units d\rrInE'jre ne\t geleral neettnS ln january I987. e 1n