rr-- Chakratit Dhanagom hongthai Varachai Instittrte for Population and Buppha Sirirassamee ph.d SociaI Ilesearch Ilead, Gradrrate SLudies l'rogrant

Transcription

1 !gmv

2 Varachai rr-- hongthai Buppha Sirirassamee ph.d Chakratit Dhanagom Ilead, Gradrrate SLudies l'rogrant Instittrte for Population and SociaI Ilesearch

3 Thesis entitled I{as submitted to the Faculty of Graduate Studies Mahidol Un iversity for the Master of Arts on February 5',1990 Jegatheswaran }"tr'lt - 8r'- ;;;;;;/ ";;;; ;;;;- (.Minor Advisor ) - ;;:;- Dh:rnagom isor ) DirecLor of Itrstitute PopulaLion and SociaI Research IIahidoI Universitl' M<>nthree Chtrl.asar.m;ry:r. M.D., Ph.l) Dean, Facult-y of GradrraLe Sturlies!t{:.rh i clo I Un i v<: rs i t-y

4 I woufd like to express my sincere gratitude to Dr. Varachai Thongthai my Major Supervisor, for his instruction, guidance and support throughout ny study program i.ncluding the completion of this Thesis. Deep appreciation is also exlend to Dr. Buppha Sirirassamee and Mr. Chakratit Dhanagora my Minor Supervisors, for their insightful comments anil suggestions. '. My.iriends and colleaguesi ny special thanks for their valuable advice and comments. I also tender my thanks to the Royal Thai Government and World Health Organization (WHO), fortfinancial support. FinaIIy my farnily, especially my wife, for her support and understanding throughout the years of my education, I am greatly indebted, and my beloved daughter Sharnila Loo hereby being acknowledged. Sandrasclieram J egaltlre swa ran

5 Name: Oate of Birth: Place of Birth: fnstitutions Attended: EDP ComPuter Institute Colombo, Sri Lanka. June 1986 Sandrasekeram Jegatheswaran January 30,.1952 Batticaloa, Sri Lanka Certificate of ComPuter Prograrring (COSOL and BASIC) UniversitY Of Coloubo Sri tanka t DemograPhic and Training Institute O"""rU"" 1980 Post- Grad' Diploma in PoPulation Studies UniversitY Of Colombo Sri Lanka Faculty Of Social erol sciences ;;;;;t." 1e?8 special Desree in DeveloPnent Stddies (B.Dev.S).

6 Thesis Title N"r" Degree Thesis Supervisory. Committee vii Proxinate Determinants of Fertility' in Sri Lanka: An Analysis of BonEiaarts' Model Sandrasekeram Jegatheswaran Master of Arts in Poptilition and Social Research Varachai Thongthai ' Ph. D Buppha Sirirassamee ' Ph. D Chakratit Dhanagom' M P A Date of Graduation 5. February B.E (1990) ABSTRACT Pr^xi mate deter-i nrnt s i:: f e:r+-i 1i +-y heve bcconc increasingly important and of much concern to policy makers, prolran planners and nanagers in mdny changing societies of the Asian region.,. The objective of this study is to determine the effect of the proximate variables to fertility in Sri Lanka 19?5 and 1987' and to assess the trends in the proximate determinants of fertility by geographical sub - groups. fn this study, (applying Bongaarts' model) the three principal internediate variables are considered inhibitors of fertilityr because fertility is lower than its naximum value as a result of delayed marriage, the use of contraception and postpartum infecundability induced by breastfeeding.

7 .: InSriLankaatwo.intermediatefertilityvariables areresponsibleformostofthevariationinfertititylevels of popuiatiods. Those are nameiy contraception and proportions of narriage - are the most important determinants offertility.highprevalenceofsterilizationand traditional methods has a strong fertility in different geographical reglions in Sri Lanka' Fertility viii inhibiting effect Thisstudyisbasedondata-fromtheSriLankaWorld Survey (slwfs) 1975, and Sri Lanka Demographic and Health Survey (SLDHS) carried out in 198?'

8 1.1 ImPortance of the study" 1.2 Study obiectives. 1.3 Review of literature' ' ' ' ' ' ' ProPortions of marriage I ContracePtion '4 Postpartum infecundabiiitv' " ". " "' 1.4 ConcePLual framework 1'5 organize of the study' 2,1 Sri Lanka World Fertilitv Survey (SLWFS) ' Sri Lanka Demographic and HeaILh Survey ( SLDHS ), and Study Area" 2,2 MethodologY 2.2,1 fndex of marriage 2.2,2 fndex of contraception' " ' CIIAPTER 2: SOTTRCES OF DATA AND STUDY AREA' METHODOTOGY

9

10 Page Age specific fertility rate and total fertility rate for 1975 and I 981 (SLWFS, Census)... " " Proportion of women married, Sri Lanka f"t'fgzs and 1981 (SLWFS, Census)" " " 40 [It"f-marital fertilitv rate for 1975 ""J-igar (slwfs, census).. " " " "" "' 42 Breastfeeding and PostPartum infecundability for 1975 and 1987 iiiwrs, SLDHS): Percentage of contraceptive used, contraceptit effectiveness, index of " ;;;a;;""itiott for 19?5 (sl,wfs)""""" 48 Percenta$e of contraceptive used, contraceptire effectiveness, index of contraceptiot for 198? (SLDHS) " " 49 fndexes of proximate determinants for Sri Lanka by Zones for 19?5 (slwfs) " " ' 52 Indexes of proximate determinants for Sri Lanka by Zones for 198? (SLDHS) " " ' 53 Summary of fertility inhibiting effects of the proximate determinants Sri Lanka Uv 2"""= for 19?5 and 198? (SLWFS' SLDHS)' 55 Percentage of contraceptive uset index of marriig", index of contraceptionr &nd index of postpartum infecundability for Sri Lanka 19?5 and 198?."".."""' 60

11 Population, Zones area and density of Singulate mean age at marriage by Sex and Zones, 1953, lg7l and 1gg Live birth rates, death rates, infant mortality rates and maternal mortality rates by Zones 1gZ Percent of eligible respondents who were current, contraceptive users at the time of survey, by methods, Sri Lanka. 1975, L and 1gg?,... Percent. married women using of contraception for 1975 and 1gB? (stwfs, sldhs)...;...r... women and type using for rates of and over

12 LIST OF FIGURES Figure 1.1 RelationshiPs between inhibiting effects o variables and various fertility the fertilitythe intermediate measures of aaaaa Page 4.L Shares of the difference between actual and potential fertilitv attributed to delayed marriage, contraception and post -partum effects by the Proximate determinants of fertility model, by selected Zones (19?5) and adjusted Zones ( 1987 ) ' ' ' ' ' ' ' ' ' ' '

13 xrv Map 2,t 2.2 2,3 Zones of Zones of Adjusted Pagc Sri Lanka 19?5, SLWFS 26 Sri Lanka 198? ' SLDHS 26 Zones of Sri Lanka

14 CHAPTEB 1 INTRODUCTION 'Among the countries of-south Asia' Sri Lanka' with a biith rare ;;-ra-;;"'iooo, has achieved bv far the lowest tertiiiti- ievel. The sri Lankan fertilitv transition has-uirusual aspects and certain puzzling features that have lot - -b""" fully explained" icafaweff et al, 198?:1)' The decline of mortality in Sri Lanka started as 1920's and paved the for a subsequent early as the paveo tne way 'L L'I q - qv decline in fertitity. In the 1920'1 the mortality rate _infant was around 190 deaths Per 10QO births. By the late 1940's it had fallen to around 100. After the eradication of malaria after the second world war, the infant mortality rate fell even more rapidly. By , it*stood at 39' one of the lowest in the developing countries of the region (Nadaraiah' 19?6; United Nations ' 1986 ) ' Althotrghthedeclineofmortalitybeganinthe 1920,s, fertility did not begin its decline until after the Second l'lorld War, and even then fertility declined slowly ' During the period the crude birth rate feil gradually,fromaround39toaround36'inthe1960'sit ' Ig74' After 1974 it began a rap.id decline,. reaching 27 tn fl'uctuated between 27. and 29 until The total fertilitv : rate feii from 5.0 in 1961 to 3'4 in ' to 2'8 in 1 98?

15 Muchofthedeclineinferti}ityduringthisperiod a ^^^-l: #a -d was due to the postponement of marriage. According to varrous censuses, the (singulate) mean age at first 2 marriage for women rose by three years since the end of the second world Wart making it one of the highest in the region (table A.2, Whilethisrise.intheageat-marriagewasonecause of the decline of fertility, also declined owing to the use of contraception. The use of contraception was not measured until 19?5, but many of the older couples interviewed that. year were aware of contraceptivelmethods like "coitus interruptus" and the 'lsafe period method";, (vaessen 1980) which are still widely uied throughout't'he country' At the time of the 19?5 survey, 32 percent of married women agdd were practicing contraception' By L982 this percentage had risen to as high as 55* percent and by 198? it had risen to 61 Percent. study is to bring together the available Iatest information on sri Lanka to demonstrate that differences in fertility among Populations are largely due to variations in only three intermediate variables. rn addition, estimates of the fertirity effect and trends of these factors a.nd of the IeveIs of general marital natural

16 ,/ii., fertilitywillbemadefordifferentzones*ofsritanka' using sri Lanka world Fertility survey (sli{lfs) 19?5' and sri Lanka Demographic and'health survey (sldhs)' 1987.(Maps 2'L and The ImPortance of, the study Like other developing counlries in southeast Alia' rapidpopulationgrowthisaseriousproblemforsrilanka. r rl ^ Thepopulationproblemwasrecognizedandthegovernmentlaid down a comprehensive and specific popuration poricy aimed at contr.olling popr.rlation growth in 19?8 r The majtir objectives ofthepo}icyweretoreducepopulationgrowthandtoinform and provide family planning services throughout the country' The steps taken by the $overnment included the assignment of the responsibility to the ltinistry lf plan rmplementation, Paymentofacashincentivetoacceptorsofpermanent methods of family pranning, appointing private dearers to seii contraceptives (ora1 piils and condoms), coordination of allfamitryplanningactivitiesofthegovernmentandnon. governmentagencies,obtaininbassistanceofinternational organizat"ions,enhancingthemedicalandclinicalservices' and established a communication system for family planning' *Note: - fn this study bhe term Zone is oniy indicates the sub groltps. of Sri Lanka' is not emphasize socio - economic, ethnic, cultural admini stration, ( Fami IY pi"n"i"g ' services/campaign, supply etc ) Ianguage etc ' '

17 But, still at least half of all stiil Practiced by neans other national faurily planning Proltram fertility control there is than those offered by the (SLDHS, 1987).4 Over the Past decade' fertility declining renarkably, representine one Sri Lanka has been the nost dranatic decreases recorded for any developing country (caldwell et al 1987). During the 1960's the PostPonenent of Earriage accounted for half of the decline in fertilitv (Uttit"d Nations, 1986). The 1985 SLCPS confirns the high reported prevalence in the 1982 SLCPS of the use of traditional contraception and sterilization in Sri Lanka and provides some of the Daior reasons for these contraceptive use petterns. Use of both rhythn and withdrawal is widespread and has contributed to the significant decline in Sri Lankan t fertility in recent years (SLCPS 1985)' Modern teuporary ' rnethods such as oral contraceptives, IUDts and injectables appear to be not readily accepted by nany Sri Lankan couples (slcps, 1985). Relation to most other countries the use of ' traditional nethods in Sri Lanka is quit hi6h' In a study of contraceptive use patterns in 42 developing countries where prevalence data are available, it was found that only Peru showed a higher prevalence of rhythm use than Sri Lanka' with L7%ofallcurrentlymarriedPeruvianwomenusingrhythmin 1981, compared to t3% of Sri Lankan women' Almost one- fourth ofallconbraceptorsinsrilankawereuslngtherhythm

18 nethod in 1982, according to the 1982 SLCPS' DespitethefactthatfarnilyPlanninE hashadauajor impact in reducinel fertility tn Sri Lanka (Caldwel et air 198?) ' the use of nodern contraceptives to delay the first birth is very low' This Patterns reflects the policy of the fanily planning Program which enphasizes the use of modern.. c.ontraceptives such as.the condon' IUD' piii' injection' and sterilization. It appears that a Possible next step in population policy is to focus nore at the beginninb of child bearing, bnd emphasizd the use of contrsceptives as a uethod of spacing as we'll as liniting fertilitv'. The adoption of' rhythn by a non-catholic population has resulted, usually in coniunction with withdrawalr in very effective use. Rhythn has undoubtedlv played a central role in the Sri tankan fertility decline and it is sure to be a significant element in fertility regulation for a long tine to come. Most PubIic health personnel recognize this and nake IittIe attempt to convert PeoPIe who are satisfied with rhythm to those methods often described by fanily Planning programs as "effective methods" (Caldwell ' 198?)' As a result of this patternl traditional methods is thekeydeterminantofovera]}cont.raceptivepractice'its needfurtherresearchandanalysis'moreontotheuse,and use-effectiveness of Lraditional methods such as withdrawal and rhythm in order Lo illustrate. the contribution to overall i""tiiitv regulation (SLCPS ' 1985)'

19 Despite the fact that famiry pranning has had a major inpact in reducing fertility in Sri Lanka, the propoitions of marriage, postpartum infecundability and abortion' explain 95 percentofthevarianceinthetota}fertilityrateinsri Lanka in t9?5 (Bongaarts' 1982) ' Without an adequate understanding of the effects of Proximate variables ( proportions of marriage r contraception tpar ' infecundability and abortion) on total fertility rate in sri LankarPoliciescannotbecarriedoutSuccessfully.Astudy oftheseproximateferti1ityvariabiesandanev11uation the contribution of ea.ct-t to fertility itself t=" lherefore' essential if ' understand fertili'ty behavior" L.2 StudY Objectives The obiectives of this =tud) are to determine the effectofproximatedeternrinantstoexplorereducingof fertilityinsrilankafromthelgtssrilankaworld FertilitySurvey(SLWFS)andSriLankaDemographicandHealth Survey (SLDHS) 1987 by Zones' This study therefore aims to provide: - (1). To determine the variables to fertility' in Sri Lanka, 1975 and 1987 'effect of the Proximate

20 (2) ' ro dete rmi nants deterninant s assess the 'of fertility of fertility.7 trends in the Proxinate in Sri Lanka by Zones and decline during ' 1.3 Review of Literature The cons i dered study are Abort ion.bonelaarts ( 1982 a), has developed a model to tleasure the effects of a number of individual bbhavioral patterns on fertility. The nodel allows the proximate determinants to be deconposed thereby allowing a better understandine of the 't causes leading to changes in fertility' If a Proxinate determinant changes (such as contfaceptive use or narriage patterns) then fertility necessarily changes also' provided that other proximate determinants remain constant' eventhough it is not always due to changes in socio - economic d.eterminant maior proximate variablesr which will be in connection with objectives of this (1) Proportions of marriage (2) Contraception (3) g eneral i y not necessary to deviate the same e f fort to analyzing and measuring each of Lhe proximale determinants because they are not of equal interest' Two criteria can be applied to seiecl the intermediate variables that deserve most attention. The first is the sensitivitv of the fertilitv

21 ' rate to variatlon in a deterrninant; it is relatively uninieresting if large variation in it pr,oduces only a mj'nor change in fertility' The second criterion is the extent of a determinant's variability amo[g PoPulations or over tine' A relatively stable determinant can contribule littie to explaining trends or differentials.and hence is less inportant' The overall rating ' based on both criteria' indicates that four proximate determinantsnarriage' PostParturu infecundability': cbntraception' and induced abortion - ard the rnost idportant for the analysis of fertility leveis and trends (Bongaarts' 1982 a)' ' '. Sndll number of intermediate fertility variables are ' responsible for nost o'f the variation in fer'tili'ty levels of populations. For intermediate factors - proportion married' contraception, induced abortiln and postpartun ' rin f ecu.r-rdabi 1i ty - are the rirost important deterllinants of fertility. These four factors explain 96 percent of the varianceinthetotalfertilityrateinasampleof4l populations that included developing and rleveloped countries aswellashistoricalpopulations(bongaarts,l982b) ProPortions of uarriage In Lhe least developed countries of the Third worldr ttre mean age at first marriage is typically between ages fifteen and twenty ' t,iith increases in lileracy and school enrollment, changing roles of women' and other social

22 ,9 and' economic developments ' age at marriage rises ' As consequence'themeanageat,marriage.inasubstantia}number of developing countries now slightly exceeds twenty years of a PoPulation is female age at marriage has important for the study of fertility, becagse the length ofthefemalereproductionperiodislargelydeterminedby the age at which a woman enters marital union (Mehtab' 1979) ' narital fertility in Sri Lanka (Littte "td Perera' 1981)' A residual effect of age at marriage Persists after all socio '- dconomic factors in the study have been controlled'!r indicating that its influence cannot be attributed solely to the higher socio.- economie status of women wh9 marry late' ^A t - : - d-: r ^-tra olrarrf h, Cline in fn 1960's in Sri Lanka about half of the de the total fertility rate was due to declining proportions married. At the national levei, the totar fertilit:r rate ferr from 5.0 to 4.1 children per woman, 0.5 of which was due tochangesintheproportionsmarried.theextentofthe,, contribution of changing pattbrns of marriage on declines in thetotalfert,ilityratevariedfrom3gtottpercent. changles in the proportions married affected virtually whole country, but the strongest effects were felt in the eastern and northern districts (Zones 8 and 9). (united the

23 In Sri Lanka during the intercensal period ' the sharp rise in the sin8ulate nean age at uarriage for women fron 23.5 years in 19?1 to 25'1 in 19?5' and its decline to 24.7 in 1981' For men it dropped slightly fron 28.0 year3 in 19?1 to 2'l '9' in 1981' The sex iurbalance at the marriageable ages in 1971 and 1981 due'to the shortage of men, who employed in the Middle East (Fernando' 1985)' Derived from Snitht s study focusin6 on Asian nuptiality, ttlarriaele behavior in Asia is assumed to be affected by three key modernization Processes:- educationt urbanization, and the expansion of non- agricultural employnent (Snith, 1980) ' The Positive relationship between female enployment and postponenent oft Eerriage was confirned in his study of eighteen'asian counlries' Snith (1980) found that societies with high labor force participation rates tend to have a high rate of late narriage' Education is also found to be highly related to the postponenenl of marriage. In Southeast Asian countries ' the greatest change in aeie at marriage' is associated with a changeinthefbnaleliteracylevel(snith,1980)..education has both direct and indirect effects on age at first marriage. DirectIy, higher education means a Ionger time spent in school, thus later entrance to the labor force' and Iater marriage. Indirectly, edrrcation serves to provide

24 better job.opportunities for without depending upon their al, 1980). 11 wonen to Pursue their own goals fanily or husband (Bindfuss et t.3.2 ContracePtion contraceptive use is atways higli when wouen who have achieved their desired fanily size ' Findings froo intdinational research are c on31s Eer originally derived from developed countries data (Ross' Sivin t Kreager t 1977 i ESCAP' 1982). In Taiwan, for examplet l'omen who have achieved their desired fanily size are nore effective contraceptive userst than rdomen who intend to hawe more chi Idren (Hernalin and Chow, 1971; Moots and Chang' 19?5)' In the Philippines' women who have achievet their desired farnily size are nore'iikely to adopt contracepti'on than other women {1,"i.16 et ai, 19?4) and nore likely time. to continue use over In Sri Lanka the choice and Lhe use of f arnilv planning methods are voluntary and the methods freely available include condoms, pi11s, IUD and sterilization (both ma 1e and f emale ). I ni ectable contracept ive are also available but are restricted in use owing Lo high cosl' EducationandincomeclearlyinfluencedtheacQeptance' continuation, and discontinualion of family planning practice. The f arnily planning program of Sri Lanka has

25 I( developed strategies to facilitate acceptan ce and continuation of faniry pranning practice at the lever of policyandimplementationinthefield.animportantpolicy oftheprogramisthatthechoiceoffaui}yplanningmethods andtheirusearepurelyvoluntary(unitednations,1982). InSriLankatraditionalmethodswereemployed efficiently and. their high leve} of continued use arises from '' to the PilI and IUD; based upon strong cultural res.istance locai interpretation of how these nethods function (Caldwell etair1987).srilankawillfollowthechangesinmany western countries when modern contraception has increasing displaced the natural methods ' Yet ' one should not underestimate iust how deep t2 seated is the feeling against mueh modern contraception. This is particurarry the case with \ the pill, where the opposi'tion is based not oniy on a fear of r -^ao Lrrri An rstem of ideas that its unnaturaf'ness but on an older sj explainshowitworksandtheharmthatitcando.these ideashavekeptpilluseataverylowlevel,andanyassault intermsofpublicitybythehealthserviceonsuchconcepts would mean an attack on the whole cultural notion of "hot and coid,, - concepts that are stiii of great tmp.,ortanc.e in South Asia. believe, tthot the and cold"- concepts focus that' the Sri Lankans PiII has a heating effect and that this heat destroys the sperm. This is related to South Asian beliefs in the balance of hot and cold influences, especially in food

26 ind in the danger of hot substances being doninant' It is believed that the heat not only destroys the spern but can and is likely to have a bioader inpaot dry or wither the womb ancl 1s raxerv urj on 8 wonen's health' A typical comment was ' rrthe PiIl is heating and, if you take it' you have to be able to afford coconut nilk every day" ' This comment relates to balancing a hot influence with. a cold food; in Sri Lanka, the Products of the coconut are thought to be especially cooling (Caldwelt et ai' 1987). While nodern contraceptive nethods are widely known in Sri Lanka, they are still 13 used by only a ninority of the wouen exposed to the risk of pregnancy (Inmerwahert 1981)' There hss been considerable use of traditional methods' and to a considerable extent there has been shifting fron traditionaltomodernmethods'indicatingthatsoneofthe modern nethods use is a substitution for previous use of traditionalmethods.theyoungercohortshavemuchless relative use of traditional nethods than the older cohorts' His analysis shows further that' two-thirds of currentlv married women inlerviewed in the SL!{FS who had ever used or practiced one of the najol modern methods of contraception (sterilization, PiIl' IUD' and condbm) had first used any of' these nelhods between 19?1 and 1975' Fenale sterilization the most Prevalent melhod in use at' Lhe time of the SLWFS and, of all sterilized women, over half had had the operation in the three years preceding the survey' was

27 1.3.3 Abortion fnduced abortion is one of the otdest forms of fertility regulationt it is also one of the tlost controversial. In nany Countries, abortion is a traditional nethodoffertilityregulation,especiallvwherenodern contraceptives are only gradually beconing accessible (David' 1983)."'Howevir, evidence strongly suggests that nearly resort ib abortioh to ever;nrhere, wonen'of all backgrounds sorne extent' regardless of legal codes' religious sanctionst or personal dangers ' The sitiration in India is particularly noteworthy because of the conparatively low incidence of legal abortion since the Medical Termination of Pregnancy Act nade IegaI abortion available on request in 19?l';'.The reported number of Iegal terrnination lncreased fron 24'000 in 1972 t'o 278'000 in 19??r while the nunber of illegal abortions was estimated at between four and six million (David' 1983) ' In countries where abortion and modern contraceptives' rnt to learn why some are equally available t it is important to learn wuj' s r{ome(t bontinue Lo 'reiy on abortion rather than on contraception. The question.arises as to why wonen wait' until pregnancy and resort to the extreme step of seeking abortion quacks, thus risking their Iives' by unscrupulous doctors and instead of accepting modern f983 ). The explanation is contraceptive technology ( Pieris ' clear for the most conservative

28 . areas. There, women wourd hesitate even to purchase pirrs or eondoms in the neighborhood, even if freely available, for Iocalgossipissuchthatimputationsofpromiscuo'l'""* relations mlght go the rounds of the community and tarnish thefamilyreputation.thosewhoseekabortionsarefromthe upper and middle classes who try to limit their families to 3 or 4 chirdren. rt has been proposed that abortion be l.egalizedlandthat.thiswil}allowon}ywhatisnolf.done iliega1iy to be performed 1ega11y by qualified (Pieris, 1983). 15 persons Postpartum infecundability BreastfeedingplaysanimportantroleindeveloPing countries because of its relationship with child health and \ birth spacing. rt is weli- known that breastfeeding has a significantimpactonreducingmorta}ityininfants.aside from the maior role, breastfeeding is equally important in controllingfertilityindevelopingcountries.postpartum infecundability associated with the practice of breastfeeding isamajordeterminantofspacingbetweenbirths,with resulting reductions in overall fertility Ievels (Haffman' 1e84).

29 ,16 the nedian durations of breastfeeding are relatively long or the rates of contraceptive use are relatively highr median birth intervals are longer than in countries where the median breastfeedin6 durations are shorter and contraceptive use is less widespread. In countries where fertility is noderate or high, however, the incidence of very short birth intervals is higher when nedian breastfeeding durations 8re relatively shoit. sogre.;analysis shows that...a. strong, independent, and consistent negative relationship between amenorrhea and contraceptive use (Lankaran and Wanikoff-ts' 1985)' Wonen who were less than four months Postpartum were also less likely to use contraceptives' They found a negative relationship between breastfeeding and contraceptive use for users of oral contraceptives. It shows' it is possible that wonen in the innediate.postpartunperiod'esllteciallythose.whoare Iactating and amenorrheic, are noi as highty notivated to use contraception as had been supposed' Proximate determinants are the biological and behavioral factors through which socio - economic and environmenta.i variables affecting fertility' among societies' the level of fertility varies due to variation in marriage patterns and in the fevel of natural marital fertility' Postponement of marriage and marital distruption shorten the Iength of exposure to childbearing and result in a level of fertility below the theoretical maximum IeveI of natural marital ferlilily. Natural marital fertility is affected bv

30 postpartum infecundability' waiting time to conception' In a controlledfertilitysituation,deliberatecontrolofmarital fertilityiscarriedoutthroughtheuseofcontraceptionor the practice of induced abontion' l7 1.4 ConcePtual framework developed. an aggregate fertility model that describes the relationship between fertility the proximate determinants' Figure 1.1 iltustrates that the Proximate determinants influence the level of fertitity in a society' if atl the proximate determinants are operating the Ieve1 of fertility and 'cou].d be as low as the level indicated by TFRI If however,allofthewomenmarriedatrsyears(nodelayed marriage), the level of fertility would increase to the level indicatedbytm:movingfurtherifnocontraceptionareused andinducedabortionisnotpracticed,theleve}offertility wouldrisetothelevelofthenaturalmaritalfertility breastfeedingandpostparlumabstinence'thelevel-of fertilitywouldfurtherincreasetotf'thetotalfecundity rate. This study utilizes mainiy on ly three types of variables. The first comprises marriage second comprises '

31 Figure 1.1 Relationships between the ;ii;;i" of the internediate variables various measures of fertility fertility-inhibiting and Tota1 fecundity rate Total natural marital fertility rate Total marital fertility rate Total fertility rate t'[ l rnr l lt tl t! ll "+---- i lr ll FertiIitY-inhibiting effect of: PostPartum infecundabi I ity ContracePtion and induced abortion Marriage u j Source: Bongaarts ( 1982)

32 contraception variables are as a dependent and third treated as variables. postpartum independent labor The force on induced abortion in computation of Ca is and fenale In order to achieve the obiectives' this study analysis is carried out at the level of aggregates. Estimates arepresentedatnationallevel,andsubgroups(zones).the strategyadoptedinthisstudy.isbuiltexplicitlyonan I approach to fefti'lity based on thre Plo,cess of total fertility; Having attempted a systemitic description of the role played by the proximate determinants at, each stage in a women,s reproductive Iifetime, this study will tried to relate these pieces of information to overall total fertility Ievels. An attempt has been made to assess the relative role ofeachofthemainintermediatefertilit,yvariablesto overall 'fertitrit,y levels ' In order to archieve the obiectives a systematic variable indicators has been derived for each proximate variables. (See details in chapter 2)'

33 Organize of the study ThestudyisorganizedinSchapters.Chapterlis thelntroduction.chapter2describesthemethodologyand sourcesofdata.backgroundcharacteristics,patterns,trends ofthemajorproximatevariablestofertilityinsrilanka are discussed in chapter 3. rn chapter 4 irrustrates effects andtrendsin:h.proximatedeterminantsoffertilityand chapter 5 is summary and conclu5ion''

34 2.L Sri Lanka 'tlorld Feitilitv Survey -' : i;"t" Demolraphic cnd Health Survey study area This study is based on data from the SLWFS 1975 and SLDHS 198?. In SLWFS 19?5 interviews were conducted f roro August to Octoberr 19?5' Interview was conducted with each etigible woman in the household where eligibility was defined as ever married and between 12 and 49 years of age' THe Sri Lanka fertility survey was the first in\ a series of surveys designed to Provide the governnent Lith up-to-date and meaning fui data required for effective Planning and monitoring of its social Programme' In SLDHS the target population was defined as 6r170 eligible respondents were identified and interviews were completed among 5,8'65 ever ruarried women between age group for a response rale oi 95'l percent' The Sri Lanka Demographic and Health Survey (SLDHS) is an important link in a chain of surveys carried out in Sri Lanka in the past decade or so. Having been designed a's part of and

35 22 international survey progrin and rnodeled on the lines of the trends over tise within Sri Lanke es well as for cross ecological criteria' nine zones were created' Accordingtothesll{Fs'forthepurPoseofsampling' t{orldfertilitysurveyprogranlthesldllsprovidesand exceptionally valuable source of data for the estination of national comparison' on the basis of socio-economic and the country was divided into 6 zones of sonewhat similar socio - economic characteristics ' The zones are located on a nap of Sri Lanka in Map 2'1' As the citv of Colonbo raetropolitan area ( zone 1), the South- i{estern Iowlands (zone 2lr the South - Eastern Coastal bett (zone 4)r the Northern peninsula and adjoining districts ( zone 5)' South Central Hil1y areas (zone 6)t and other areas (zone 3) (see Map 2. 1). The citv of Colombo' the maior urban center of the country' forned zone 1, and is a mixture of alurost all ethnicr religious and social groups ' Zone 2 consisted of the South- western Iowlands where the maiority of the population are Sinhalese Buddhists and Christians' It is an area with a relatively high Ievel of urbanization and its proximity to' Coloinbo makes it susceptible Lo the diffused effects of major developnents in the city' The dry South - Eastern coastal beltformedzone4,thirtypercen!ofcountryismoor populaeion live in this region and constitute about one third of ils population' The maiority' however are Sri Lankan

36 Zone r-colonbometropol'itanareaconsistingofslwfszone 1 and Parts gf zone 2'. Zone 2 - Colombo feeder areas arid Northern part of SLWFS zone 2 Zone 3 - South western coastal Low lands corresponding to Soutn'ern Part o f SLWFS zone 2 ' Zone 4 - Lower South Central hili country corresponding Lo western and Southern part of SLWFS zone 6' "*"t rii.j"ji=jti"i" t concentration' of estates 'ftn.

37 z4 ZoneS-SouthCentralhitlcountrycorresPondingtopartof SLWFS;;;"Swithaconcentrationofestates',eLLs v "-ith ;l;r of minor irrigation schemes. Zone 7 - Rain fed D'ry zone coverin$ the rest of SLWFS zone 3 ' ZoneSEasterncoastalb.elt'correspondingtoSLWFSzone..4 i;;a included in 'SLDHS ) ' zone g Northern province corresponding to slwfs zone 5 Although the survey originally planned to conduct interviews in arl nine zones, civil disturbances in zones I ands(thenorthernandeasternprovinces)prevented interviews from being conducted there' These zonest which contain approximately L4 percent of the 1986 estimated populationofsrilanka,havebeenexcludedfromsldhsl9st. Althoughsmall,thecountryshowsagbographical diversitywhichhasledtotheconcentrationofminority ethnicgroupsinparticularareasofthecountry.forthe samplingrtherefore,thecountrywasdividedinto6zonesin lgtsandgzonesinlgstofsomewhatsimilarsocioeconomic characteristics.forthepttrposeofanalyzing,thecountryin lg87 has been divided into four Zones (IIap 2'3) of somewhat similar geographical sub-groups as in 1975!

38 ce located on a map of Sri,.r::.;:"::."',-,,," rooserv described (slwfs 1975) Zone 6: Pollonaruwa Matara Moneragalle Hambantota Zorrg 1:Colombo CitY (A).Zone 1:Colombo city Zone 2:GinPaha (B).Zone 2:Gampaha Kalutura Kalutura..' " Gatie (stdhs 1987) : Zone 3:Gafle (F). Zone 6 :Trincomalee Batticaloa AmPara Zone 7:AnuradhaPura Puttalun Zone 8 : Trincomal'ee Batticaloa AmPara Zone 9:Jaffna Vavunia t Mannar Mullaitivu

39 ltap 2.1

40 This study analyse separately the 19?5 and 198?' and MaPs 2'1 and2.2wiiibeused(chapter3).butwhencomparingbetvleen. 19?5and198?theareascomParewilluseZoneA-DasinMaP 2,1 and 2,3 elc. (chapt'er 4)' MethodologY Bongaarts (f982 b), has developed a simple nodel for assessing the impact of specific internediate variables on fertitity in any population ' The raodel includes a set of "indexes". Each one indicates the extent to which fertility is reduced from rnaximat levels by a specific intermediate variables (a "proxinate 'determinant" '' ad he also terms then) ' The index value actually represents the fertility expected under observed levels of the intermediate variable' as a proportion of fertility in the absence of any fertility reducinel inpact of the variablel and thus 1'00 ninus the index value is the estirnated Percentagle reduction in fertility due to the intertnediate variable' The indexes provide a sunnary measure of the. role' of each var'idbie' and have the convenient ProPerty of suronarizing the inpaet of ai1 of the internediate variables in 'the sane metric - namely' proportionate fertility reduction - and thus Pernit conparison of the relative inportance of different' internediate variables' Multiplication of two indexes Yields an estimate of the combined effect of two ini'ermediate variables (Singh and Casterline' 1985)

41 fn the decomposition: - TFR(TotaIFertilitvRate)rTMt'ott'MaritaI Fertility Rate), TN (Total Natural Fertilitv Rate) and TF (Total Fecundity Rate) are relate to the indexes of cm (index of proportions of marriage), Cc (index of contraception), Ci (index of Postpartum infecundabilitv)' ald

42 Cn' Cc' and Ci measure the fertility - inhibiting effect of the. three nain internediate fertility variables: Marriage' ContracePtion' and PostPartum period' Cm = Index of marriage ( equals f. i'f -'I1 reproductivt ars married and of rnsrriage "*gt ) ' "ui.".. Cc = Index of contraception of contraception and Percent effective Ci = Index of postpartum (equals 1 in ' infecundabilitv th" lacta-tion and postpartum austinerije'-;j "bj;;;;-;i 0 i.f the duration of infecundability is infinite) ' TF is, the. total Potential fertilitv ( total fecundity), or the level of total fertility 29 one might expect ifallwomenwerenarriedthroughoutthereproductiveae e range' if there was no use of contraception and abortion' and if the post - Partuto period $as not extended by lactation and abstinence. wonen of 0 in the (equals 1 in the absence if all fecund women use contraception). Each ' of ' the.index.es 'can!e calculated f r<in neasurenents of the intermediate fertility variablest as rs illustrated belolr: - 2.?. Index.of DarriaEle Cn is an index of reflecting the relative loss of the fact that nost.h'omen ere not the inpact of narriage I potential fertility due to continuously narried between

43 the ages of 15 and 49 Years' ttis is often calculated as: Cm= TFR Zzt(a\ (A) Where: m(a) equals the females, by age. f(a) fertility rates. proportions currently married among is a schedule of age - specific 2,2,2 Ihdex of contraception Cc = 1. - s * e * u (B) Where +r^^ --o.,olanr c contfaceptive use u is the Prevalence of currenl among married wottr rrr is the average use effectiveness of contraceptionr- and s is a sterl,tity correction factor' T1}" r.: sterility correction factor is estimated t'o be 1'' i. f.ot runtries (fn ttri-s study 1'1 wi'11 be used to and 1.15 in developed countries (Bangaaits' 1980).

44 ,r Weassunedthefo}lowinglevelsofeffectivenesss '1 '00 for sterilization 0'90 IUDs 0.87 Pilr 0'?5 for condoms and O'?O ior all other nethods' (Bongaarts 1982 ) (e) equals the proportiona'i r'eduction in the nonthly 'of drobabiliiy conception due to the use of contraception enong fecund women (Bongaar 1980)' In this studv the ' nedium contraceptive ef-fectivenesg is taken frorn the U'N' ' slandard of effectiveness (Bongaartd' 1982)' 2,2,3 Index of postpartuu infecundability 20 ci = i (c) I{here: i is the mean duration (in nronths) of the period during which the woman is in Post-Partum amenorrhea or Postpartum abstinence' based on a sample of births' If a direct estimate of i is not available it is gossible to nake an indirect estimate from the duration of breastfeeding. The foiiowine table is based on the model relationship between lactation and infecundabi I i ty :

45 32 Mode1 estimate of Median ^ a,rei"n PostP.artun "t ' Ame '' Since differences between tite mean and median duration of postpartum amenorrhea appear to be small' this table can be used to estimate i ' except at the shortest durationsofbreastfeedingwheretheinfecundableinterval mayexceedtheamenorrheaintervalduetoabstinenceor}ow frequency of intercourse in the months iurmediately following a birth Equation(c)willbeusedtocalculateCiasmedian duration of postpartum amenorrhea is available' Sinceduetothelackofinformationoninduced abortion in the 1975 SLWFS and 198? SLDHS, the computation of Ca is omitted from this analysis. The inrlex of marriage (Ca), indexofcontraception(cc)'indexofpostpartum infecundability(ci)andtotalfertilityrate(tfr)wiiibe measured here.

46 2.3 Conclusion FertilitY q'lt0qufffvn$- This study is based on data from the Sri Lanka Wor1d SurveY L975 and Sri Lanka DernograPhic and Health Survey(19?5)andSriLankaDemographicandHealthSurvey (SLDHS) carried out.in 198?' According to 19?5 countrv has -. :- root divided been. divided into 6 Zones and in 198? it h?" been into g Zones. But for the purpose of anaryzing the country has been adjusted into 4 regions simiiar to 19?5 The study emphasis here is on estinating the proximate determinants for 1975 and 198? (adiusted Zones) ' Three main intermediate fertility variableg, namely marriage (Crn)rcontraception(Cc)randpostpartuminfecundability(Ci) will be measured here', r r I b '9+o

47 This Chapter consists of two sub sebtions' The first section describes background charactetistics of the study population. Patternrs and trends of fertility' marriage' breastfeedirig and contraception are reported in section tho. 3.1 BaekElround Characteristics of the study population ^ Since independence ' Sri Lanka has been Primarily a welfare state t uhich has enabled her to make significant advances in social developmentr of singular importance are the achievements in education and public health' sri Lanka's social ind'icators such as Iiteracy rates t infant nrortality rates' Iife expectancy and total fertility rates are far better than Dost countries in the region or: what one might expect on the basis of her national income' GeograPhicallyr the island of Sri Lanka is situated offthesoutherncoastoflndiabetweennortherllatitudess. 55' and g'so' ana eastern longitudes 75o 42' and 81o 52" The Islandr an elongated pearl in shape' spans an area of 64t651 square kilometers' stretching a naxinirn length of 435 kiloneters and a width of 255 kiloneters'

48 SriLankapossessesawealthofdemograPhicdataupon which a considerable degree of reliance can be placdd ' The main sources of population information are the periodic censuses which started in 19?l and the systen of vital registration which was established in 188?' In addition a number of sanple surveys carried out in the country in recent years provide an abundance of data on demographic, sociall and economic characteristics of the people ' Two of the most inportant of these, particulerly for purposes of comparison with SLDHS 198? and the SLWFS in 19?5' As of mid- 198?' Sri Lankahadanestinatedtotalpopulationof16.2nillion.The rate of PoPulation growth 'has fluctuated during this century' Prior to World l{ar TtJo r the growth rate was betlreen 0'9 and 1.? percent. After 1945 and the dramatic decline in the death rateduetomalariacontrolefforts'thegrowthrateroseto 2.8 percent. During the Pe'riod from to 19?1 the population growth. rate leveled off and then declined to 1'7 percent during the period fron 1971 to 1981' 35 3,2 Patterns and trends of and conttaception 3.2. I Patterns and tre.inds of f erti I ity In the first half of this centurv' Sri Lankats Crude 31"g5 p6gs (CBR) fli.rctuated between 36 to 42 per thousand' The CBB began to decline slowiy f,rom around 38'2 Per thousand fron 1953, it. reached 36 per thousand in 1960' ind 30 by the

49 36 end of the decade of the 1960's' per thousand by the mid 1970's Ievel until the beginning of the declined to 27'8 Per thousand ' between 28 to 37 births Per in births Per PoPulation It declined to around and stagnated around this 1980's. It has no$ (in 1987) The CBR by Zones ranging population in 1971 and 24 to in Table A.3 shows the live birth rates pattern for the period 19? Between 1971 and 1980 the live birth rates declined from 28.5 per thousand to 24' 1 per thousand in Zone l, 2'l.8 to 25.8 in Zone 2 and 30'8 to 28'1 in Zone 6' In Sri Lanka, the change in the sex-age structure favor a higher level of fertility. For exanpler the ProPortion of females in the childbearing age increased frorn 46'4 percent in 1963 to 52.0 percent in 1981 (United Nations, 1986)' Mortality in Sri Lanka and by Zones has continued to decline during the period 19? (Tabfe A.3). A decline in infant and maternal nortality due to the introduction of maternal and child welfare services, particularly the introduction of trained midwives (United Nations, 1986). fhe re'duction in infanl mortality neant a significant incrqase in the nunber of'. surviving children in most families. Cbnsequentlyt one of th? factors to high fertility influenced fertility (Table A.3). was renoved' Another factor ' that was the low level of uaternal nortality The total fertifity ratel as a level of fertilityr is an outcome of measure of the actual the variation in the

50 proximate determinants. fn otherwords, it is the average number of children that r.eould be born alive to a woman ( or' group of women) during her lifetime 37 if she were to Pass through ail her childbearing years conf orrning t,o the age specific fertility rates of a given year. In Sri Lanka the total fertility rate (TFR) was estimated at 5.3 in 1953, and it declined to 5.04 in 1963 which declined further to 3.5 in A further decline was arrested, and it increased to an estimated 3.3 by 1981 according to the Sri Lanka census. The TFR has declined substantially to 2.8 but it should be noted that this estirdate did not include data from the Northern and Eastern provinces. total fertility rate is calculated fron inforroation on births and women usin8 SLWFS 19?5 and Sri Lanka Cenaus 1981 (Table 3.1). The result fron the 19?5 SLWFS show a total fertility rate of 3.5 for Sri Lanka and Zone 4 have highest fertility (4,1) and Zone 1 (Colonbo metropolitan area) women have the Iowest (3.1). fertility. fertility The result frorr the 1981 Census show a tota.l rate of 2.8 for Sri Lanka;and Zone 6 (Est;te sector) and? (Dry zone) wonen have highest fertility (3.?) and Zone I wonen havi the lowest fertility the TFR declined fron 3.5 to 3.3. Between 1975 and

51 specific fertility. rate and total fertility rate tg-gi (slwfs, census) Age of women 25-2.s TFR 148 rg2 I?o tl? 36 L L22 Zone lg 155 LhO 65 gz 2 2'6 zone z ii 158 ie ' 1 zone 3 26 t44 igs L '4 zone 4 ia tlg igo '2 zone 5 24 t9? ios '3 zone 6 si LLL 212 fi4 t '7 zone zo1. Iaz 82 --? '7 Source:SLWFS 19?5 and Census 1981 Note: The zorr"=."" definei aiii"""rrtly between 19?5 and 1981'

52 39 Since age specific fertility rates are free of the effect of changes in the sex-age structure of the population' they provide a norc meaningful measure in the analysis 'of fertility patterns and trends' Age specific fertility rates and the total fertility rate for Sri Lanka are given in :--- -i-^- L'ith in.reasing Table 3.1. Age specific fertitity rises with increasing age' gradually at first and then more rapidly' reaching a 1>eak in the age group ' Thereafter the rates show a downward trend in the later reproductive ages, reflecting the effect of declining fertility level in older ages'.3,2.2 Marriage A study of the rnarital status composition of the population is an irnportant aspect of demographic analysis' In societies which have not yet achieved effective 'fertility control and in which Procreation 'is almost entirely within narriage, roarital status cotdposition is a najor determinants of fertility and hence of population growth'. The singulate'mean aeie at marriage increases between 1g53 to 1gg1 about..7 years for nale and 3.5 years foi feuale. (T.able A.2). According to 1981 CensusI Sri Lanka women married at about 24'4 years and nele Darried at about 27'9 years oldl while frour 19?1 Cens:F'. averaeie age at marriage is 28.1 for nale and 23.5 for femele (Table 3'2)' In a naiority of zones the nean age at marriage lay in a narrow range :around the national average of 28'1 in 19?5 and 27 '9 in 1981'

53 zonel '081 '3?q '9?? '?e5 '851 zone 2.ogi :;;6 "991 '782 '834 zone 3.osi :iit ''u^\\ '734 '812 zone 4 'o'i :;;; :qll '832 ' '862 zone s.o16 '99? '814 '865 Zone 6.oti "i;6 'i;i '9\1 '818 '875 zone t."i :;;a '?es ' 8?1 ' 882 il;;; t slwfs 1s? 5 ' Census 19? 1 and ' - Note:the 1;il Zones ;:;ii]a-aiit'rentrv in 1e75

54 Yet' in Zone 3' 4 and 5 rural areas - wonen tended to narry at relatively younger sges' The' age at at Earriage was highes! in ttre Zone 1 (Colonbo retropo'l itan area) and Zone 2 (Southern coastal area) ' 41 > Age specific time spent within union ' are needed to. rarilal fertility rate.cglculate directly the age specific. ' 'Hoia.raa, infornation on marriage' histories are not available from the 1987 SLDHS' In this study' proportion roarried is i oqr no-o,rc aoolied essumed to be the sane as those in 1981 Census and appi to the age specific fertiiity rate in Table 3'1' to obtain the estinated age specific narital fertility rate' Thust the 1987 figures nay be over estinated Table 3'2 presents the proportion narried recorded in the 1975 SLWFS and 198?' The estimated age specific marital fertitity ratesl especially at ages 15 to 19 years, is may inaccurate due to the few cases ofevermarriedwomeninthesample.moreoverduetothe short averaie of raarital duration' the marital fertility ratesforwomenagedls-19years'therefore,donot represent the potential fertility of the whole age group' In this case Bongaarts recommends that the narital fertility rate for women aged be taken as 0'?5 of the rate for women aged (Bongaerts' 1982)' The rates for 1987, that in Zone 3 patterns of the estimated total marilal fertility the women in Sri Lanka 1975 is higher than the is 6.8 is higher (7'0) as compared to 5'3 (Table 3'3)' The women than those for the other Zones in

55 TABLE 3.3 Age specific narital rate for 19?5 and Area

56 some social 43 factors causingpeopletoearlymarriagealsoledmarriedcouplesto have more children' Their.per capita income also may be low ^-^-i n qnd tus as a result of due to low economic and social sta traditional farming systemland levels of Iiteracy rate is for males and 80 for Ifenales as compare to intermediate, 89 f mar'es arlrr ov rer!euraavu (Table A.?) Breastfeeding : The postpartum infecundability is the period during which a woman is temporarily infecund(anovulation) following child birth. The duration of postpartum infecundability and the anovulatory Period is dependent on the duration of frequentandintensebreastfeeding.ithasbeenobserved that,intheabsenceofbreastfeeding,postpartum amenorrhea normally lasts for abput two months whitst it can averagebetweenoneandtwoyearswherebreastfeedingis prolonged and intensive (Page et al, 1982)' Thedurationofpostpartuminfecundabilitycanthus, be directly calculated from the mean duration of breastfeeding based on the model relationship between lactation and infecundability from a study by Lesthaeghe and Page(Bongaarts,198o).Meandurationofbreastfeeding calculated for the purpose of comparison with findings from thelgtsslwfsandsldhslgst.amodelonthereiationof