Embed Size (px)
Citation preview
DOCUMENT RESUME
ED 070 605 SE 014 958
TITLE Studies in Family Planning, Volume 3, Number 9.INSTITUTION Population Council, New York, N.Y.PUB DATE Sep 72NOTE 28p.AVAILABLE FROM The Population Council, 245 Park Avenue, New York,
New York 10017 (Free)
EDRS PRICE MF-$0.65 HC-$3.29DESCRIPTORS *Behavioral Science Research; Contraception;
Demography; *Family Planning; *Foreign Countries;HealthiEducation; *Programs; *Reports; SocialSciences; Statistical Data
ABSTRACTIn the principal article, *A Comparison of Two
Large-Scale Studies of the Use-Effectiveness of IUDs,* the results ofevaluation of loop D use in 11 public health and hospital clinics(California Cooperative Statistical Program for Evaluation ofIntrauterine Devices) are compared with results from the CooperativeStatistical Program for Evaluation of Intrauterine Devices. Thesecond article, *Policies Affecting Population in West Africa,*examines the overall impact of governmental social and economicdevelopment policies on the rate of population growth and themovement of people from rural to urban areas. *Family PlanningAcceptors in Lagos, Nigeria,* describes a comparative study of familyplanning clinic patients. Its purpose was to look for indications ofthe kinds of persons who might attend such clinics, given furtherhealth education and the.opening of additional clinics. Three socialand demographic factors emerged from this study as predictors of useof family planning: age, parity, and educational level. The finalarticle, *Use of Oral Contraceptives in Urban, Rural, and SlumAreas,* describes a comparative study of the acceptability anduse-effectiveness of oral contraceptives in the Howrah District inIndia. (AC)
4......
U.S. DEPARTMENT OF HEALTH.EDUCATION `WELFAREOFFICE OF EDUCATION
THIS DOCUMENT HAS BEEN REPRO-DUCED EXACTLY AS RECEIVED FROMTHE PERSON DR ORGANIZATION ORIG-INATING IT. POINTS OF VIEW OR OPIN-IONS STATED DO NOT NECESSARILYuS ies in REPRESENT OFFICIAL OFFICE OF EDUCATION POSITION OR POLICY
Family Planning0,,.
VoNme 3, Number 9
A PUBLICATION OF
THEPOPULATIONCOUNCIL
A Comparison of Two Large-Scale Studies of the Use-Effectiveness of IUDs
by A. PIERCE, J. M. HILLER, and J. MCGUIRE
The results of the California Cooperative Statistical Program for Evaluation ofIntrauterine Devices and the Cooperative Statistical Program for Evaluation ofIntrauterine Devices are compared with respect to such elements as demographiccharacteristics of acceptors, continuation rates, percents lost to follow-up, andannual and cumulative rates of events by type of termination over a two-yearperiod. Altheugh differences in definitions and procedures to some extent limitedthe comparability of the two studies, the California Program's close duplication ofthe results of the Cooperative Statistical Program was sufficient to confirm thegeneral validity of the latter.
Mr. Pierce, Ph.D., is director, Mrs. Hiller, M.A., assistant director, and Mr.McGuire, Ph.D., senior statistician of the Family Planning Study of the Schoolof Public Health of the University of California at Berkeley. The research uponwhich this publication is based was performed pursuant to Contract No. PH -43-67 -1391 with the National Institutes of Health, Department of Health, Education,and Welfare, under the sponsorship of the National Institute of Child Health andHuman Development.
In this report the results of evaluationof loop D use in 11 public health and hos-pital clinks (California Cooperative Sta-tistical Program for Evaluation of Intra-uterine Devices) are compared with re-sults from the Cooperative Statistical Pro-gram for Evaluation of Intrauterine De-vices directed by Christopher Tietze,M.D., and sponsored by the PopulationCouncil. Because of the similarity andlength of the official names, the Tietze
study will be referred to henceforward asthe CSI', and the California study, whichhas beer. known informally as the FamilyPlanning Study, will be referred to as theFPS. Readers of this comparative reportare assumed to be familiar with the re-sults and methods of analysis presented inthe CSP's Ninth Progress Report (1).Until now the CSP has been the principalsource of information on the experience ofvaried categories of IUD acceptors of one
September 1972
of the major first generation types ofdevice.
Participating clinics in the FPS werebased either in county hospitals (52.0 per-cent of all first insertions) or in publichealth departments (43.9 percent of allfirst insertions). A very small proportionof first insertions (4.I percent) were per-formed in two private hospital clinics.Most of the participating clinics were un-able to sustain the burden of schedulingfollow-up more often than a month ortwo after insertion and annually there-after.
Investigators participating in the CSPincluded, in addition to public hospital-based clinics, a number of specialists inobstetrics and/or gynecology in private orgroup practice or in Planned Parenthoodor private hospital clinics. A good numberof participating physicians were deeplyinvolved in research on intrautaine con-traception; at least two of them were en-gaged in design of several types of devicesbeing studied. It is likely that, as Dr.Tietze suggested, they "took a greaterinterest in the individual woman than isusually possible in a public health set-ting." As an example of this, the general
TABLE OFCONTENTS
mg roputAnce4 Comm. INC.. 171
A Comparison of Two Large-Scale Studies of the Use-Effectivenessof IUDs
A. Pierce, J. M. Hiller, and J. McGuire
Policies Affecting Population in West AfricaDavidson R. Gwatkin
Family Planning Acceptors in Lagos, NigeriaRobert W. Morgan
Use of Oral Contraceptives in Urban, Rural, and Slum AreasMurari Majumdar, B. C. Mullick, A. Moitra, and K. 7'. Mosley
"PERMISSION TO REPRODUCE THIS COPY-RIGHTED MATERIAL HAS BEEN GRAN7E0
Hobart EllisTO ERIC AND ORGANIZATIONS OPERATINGUNDER AGREEMENTS WITH THE U.S. OFFICEOF EDUCATION. FURTHER REPRODUCTIONOUTSIDE THE ERIC SYSTEM REQUIRES PER
ANS
205
205
214
221
227
TABLE 1. Comparison of CSP and FPS number and percent of first insertions by loop D,by age, parity, and interval from lest pregnancy to insertion.
Item
Number of first insertionsof Lippes loop D
Percent of first insertionsof Lippes loop D
FPS CSP FPS CSP
Age15-24 2,233 2,831 49.0 37.525-29 1,163 2,113 25.5 28.030-34 669 1,411 14.7 18.735-49 490 1,198 10.8 15.8
Total 4,555b 7,553 100.0 100.0
ParityNone (9)a 134 (0.0)0 1.81-2 1,649 3,262 36.2 43.23-4 1,616 2,732 35.5 36.25+ 1,290 1,425 28.3 18.8
Total 4,555b 7,553 100.0 100.0
Interval fromlast pregnancyto insertion
5-12 weeks 2,301 2,301 50.5 30.5More than12 weeks 2,256 5,252c 49.5 69.5
Total 4,557b 7,553 100.0 100.0
Note. Distribution of loop D acceptors in the CSP was taken from Table 2 in C. Time and SLewit,"Evaluation of intrauterine Devices: Ninth Progress Report of the Cooperative Statistical Pro-gram." Studies in Family Planning 1, no. 55 (July 1970), p. 4.
Actually a few nulliparous acceptors received D loops, but 0 parity was defines out of the analytic popula-tion because there were too few cases on which to make reliable assessments of user experience.
b The discrepancy between the totals results from two cases with parity unrecorded, which two casesare not included in the rates analyses that follow.
c Or never pregnant.
pattern of follow-up examinations sched-uled by the clinics in the CSP in the firstyear after insertion was at approximateintervals of three months (after first orsecond menses), at six months, nine'months and 12 months (/, p. 6).
Characteristics of acceptors in the twopopulations also differ, predictably. Ingeneral, initial acceptors of loop D in theFPS were younger and of higher parity atfirst insertion than those in the CSP. Halfof the FPS acceptors of loop D were in-serted within fin .s months of last preg-nancy termination, compared with lessthan a third of the CS? acceptors. Thisimplies that a much larger proportion ofthe CSP acceptors were motivated to sc.kcontraception independently of a regu-larly scheduled "six weeks" postpartumcheckup. That level of motivation mightbe expected to be accompanied by agreater tolerance of side effects and morepersistent cooperation with research ef-forts. Table 1 compares the two popula-tions of loop D acceptors.
Definitions and ProceduresResults of the two studies can only be
compared where definitions and pro-cedures are comparable. General types ofterminating event or closure of IUD useemployed by both the FPS and the CSPare: accidental pregnancy, expulsion, re-
moval for medical reasons, removal forpersonal reasons, lost to follow-up, andreleased from follow-up. Operational defi-nitions of several of these types of termi-nation differ between the two studies.Where differences are not mentioned,commonly shared definitions may befound in the CSP Ninth Progress Report(/, p. 6). Types of Termination for whichFPS and CSP rates will be compared are:
Pregnancy (1)
ExpulsionFirst expulsionLater expulsion
Removal (2)Removal for pain and/or bleedingo"moval for other medical reasons (3)Planning pregnancyOther personal (4)
In addition, the following are discussed:
Loss to follow-up (5)
Release from follow-up (6)
Parenthetical numbers refer to notes com-paring operational definitions used inthe two studies. These notes appear here-with in numerical order.
(1) For about 37 percent of all acci-dental pregnancies recorded in the FPS,the status of the device at time of concep-tion is not known. These involve occur-rence of both a conception and either anexpulsion, a removal, or a utrAne per-
206
foration, and either the date of the associ-ated termination and the estimated date ofconception were so close that we cannotbe certain which occurred first, or one orboth dates had to be approximated, Forthe sake of scientific conservatism allthese pregnancies were charged to devicefailure. The single exception to this rulewas when the associated removal was forthe purpose of becoming pregnant and theestimated conception date was immedi-ately after the removal date. Such ter-minations were classified as "removal,planning pregnancy." The CSP followedthe same practice.
Conceptions occurring prior to inser-tion were excluded from the FPS analyticpopulation but classified as removals bythe CSP. Conceptions known to haveoccurred after another type of termina-tion were not, of course, classified as ter-minations by either study.
(2) When two or more reasons for re-moval were given, the FPS classified theremoval by the reason or combination ofreasons that appeared first on the follow-ing priority list: pain and/or bleeding;other medical; planning pregnancy; per-sonal other than planning pregnancy;other reasons (believed to be nonrelevantto device.effectiveness); reason unknown.The CSP also gave priority to medicalreasons over personal reasons for removal.
(3) "Other medical" reasons for re-moval for the two studies are listed below:
"Other medical" reasons for removalHysterectomy (completed)
PID or salpingitisDefinite or possible cancer, atypical
PAP, or for biopsyDischarge
Other genital tract pathology (cervicitis,vaginitis, and so on)
D & C or other surgery
Endometritis
Complaints of menstrual irregularity,menstrual problems other than pain/bleeding attributed to device
Discomfort of husband
Retracted threads
Anemia
Ovarian cysts
Perforation of uterus or cervixDiscomfort of user (not pain/bleeding)
Accidental removal by userCystocele, rectocele
Insertion into pregnant uterus(Excluded from FPS)
Erroneous diagnosis of pregnancy
(4) Personal reasons for removal arehandled somewhat differently by the twostudies. These removals are for "planningpregnancy" and for "other personal rea-sons" (shown separately by both studies),and for "other reasons" (originally termed"nonrelevant reasons" and including re-movals for which no reason was given).Removals for "planning pregnancy" arelikely to be coded fairly consistently be-tween the two studies; those for "otherpersonal reasons" and "other reasons"are not. Reasons classified as "other per-sonal" in one or both studies are listedon the right.
Selected reasons (OR) were screenedout of the "other personal" classificationin the FPS analysis on the groundsthat classification as defined also repre-sents unsuccessful use of intrauterine con-traceptives, whether due to psychologicalpressures or to a preference by the patientor her husband for some other contracep-tive method or for no method. Althoughthe distinction between personal reasonsrelevant to device-effectiveness and thosenonrelevant may be useful, there was nota sufficient number of these removals inthe FPS to permit reliable analysis."Other personal" and "other" reasons forremoval were calculated only for com-parison with the CSP. There is doubt asto whether they should have been sepa-rated, since in their present form, neitherthe rate of removals for "other personal"reasons nor the sum of the two categoriesin the FPS is fully comparable to the rateof removals for "other personal" reasonsin the CSP because removal at investi-gator's choice, classified by the FPS with"other" reasons for removal (OR), is ex-cluded from all rates presented by theCSP. Fortunately this category is verysmall.
(5) Loss to follow-up is defined by bothstudies as covering those women overdueby at least three months for a scheduledclink follow-up "from whom no informa-tion could.be obtained by telephone, mail,or home visit" (/, p. 6) and in addition,
Classification"Other personal" (OP) reasons for removal CSP FPS
Lack of confidence in device OP OPFear of injury (or cancer or disease) OP OPReligious reasons OP OPHusband's objection OP OPElective sterilization of husband or wife OP OPOutside doctor's advice OP OPNo reason given OP ORPatient requests pills or other contraceptive method OP OPSeparation, divorce, widowhood OP ORPatient no longer wants or needs birth control, including planned
marriage with no mention of desired pregnancy OP ORAccidental removal by patient, other nonmedical person or by
doctor after accidentally cutting threads too short OM ORRoutine removal after extended use or nonstudy clinic or PMD Removal at
policy on type or size of device investigator's ORPatient moving from area choicecop = other personal reasons; OR = other reasons; OM is a medical reason other than pain/
bleeding.
in the FPS those about whose IUD statusno information could be obtained fromhospital records.
Loss to follow-up enters into the calcu-lation of other rates by reduction of thenumber at risk subsequent to its occur-rence. Numbers and annual and cumula-tive percents of acceptors becoming lostto follow-up are shown in Table 2 forloop D, all segments, for the FPS.
Levels of loss to follow-up are generallyhigher for FPS participating clinics thanin the population of acceptors in the CSP.In the FPS, staff tracing efforts were con-centrated on acceptors who had been lostto follow-up in the first segment withintheir first year of experience. The effortswere thus allocated because of the em-phasis of the FPS on correlates of IUDeffectiveness and the anticipation that theevents of primary interest would be mani-fest sufficiently in the first year. As ithappened, some 60 percent of the firstsegment pregnancies, 81 percent of theexpulsions, and 51 percent of the re-movals for a four-year period of observa-tion occurred within the first year. It isalso true, that efforts at long term follow-up in public health clinics entail a risk ofintroducing bias through association be-tween successful tracing and characteris-tics affecting outcome. FPS staff made nocomparably intensive effort to trace ac-
TABLE 2. Numbers and annual cumulative percents of acceptors lost to follow-upfor loop D. all segments, for the FPS.
Percent of womencontinuing atbeginning ofgiven year and
Ordinal year Number becoming becoming lost toafter first lost to follow-up follow-up ininsertion in given year that year
Cumulative percent becoming lostto follow-up by the end ofspecified year of ter first inst., ticn
All Insertions First insertions
First 435 9.8Second 355 12.1Third 378 20.8Fourth 338 41.8
9.617.325.633.0
9.018.223.329.7
207
ceptors of reinsertion or terminated ac-ceptors or women with a year or more ofuninterrupted IUD use. Efforts of par-ticipating clinics to trace patients lost tofollow-up varied greatly over time andbetween clinics but were unrelated topatients' experience with the IUD.
Tietze has analyzed the first year's (allsegments) experience with loop D for agroup of ten of the CSP investigators withgood follow-up (averaging 3.3 percentlost to follow-up) and a second group often investigators with poor follow-up(averaging 13.1 percent lost to follow-up)M. The only significant difference hefound between the two groups at one yearwas in rates of removal for bleeding and/orpain. In a similar investigation using loopD first segments, FPS found that threegroups of clinics, defined by range of per-cent lost to follow-up at the end of oneyear (5.5-7.9, 8.8-9.1, and 10.6-15.1),differed in rates of removal both for painand/or bleeding and other (personal) rea-sons. One would expect that clinic prac-tices, resources, and circumstances wouldaffect both the levels of voluntary re-moval and return for follow-up. For ex-ample, where clinic staff is inadequate tothe task of contacting patients overduefor a follow-up visit, it seems likely thatpressures on staff resources would alsoinhibit the greater investment of time insupportive therapy for such side effects ofIUD use as pain or bleeding, and thus in-crease the likelihood that complaintswould be handled by removal. This wasnot tested directly, however; no informs-
Our own analysis suggests that the effectsof follow-up loss would vary considerably withmore detailed breakdowns of the data, and wewould not regard it as safe to generalize Tietze'sfindings on this point beyond the population onwhich it was based. A detailed analysis of theeffects of follow-up loss is in an early stage ofpreparation.
7).
Lion was collected on complaints inde-pendently of termination by removal.FPS data do not support the finding thatdevice-dependent clinic differences inbleeding and/or pain removals, as re-ported by Bernard from Pathfinder FundStudies (3), are independent of complete-ness of follow-up; the risks of becominglost to follow-up and terminating deviceuse are not independent of one another.To the extent that becoming lost tofollow-up is associated with characteris-tics that are in turn associated with se-lected types of terminations, the rates ofremoval for pain and/or bleeding and theother types of termination are suspect.
(6) Release from follow-up with devicein place includes for both studies the fewdeaths (from causes known to be un-related to device use) and all acaptorswithdrawn from observation with deviceknown or believed to be in place. Re-ferral to other clinics or physicians is in-cluded in this classification by the CSP,but it was not feasible in the FPS to keeptrack of these referrals, which thereforeoften were classified as lost unless subse-quently contacted. It is assumed that ac-ceptors are released from follow-up inde-pendently of their experience with theIUD or characteristics related to experi-ence, and that their being withdrawn fromthe population at risk does not introduceany systematic bias.
Neither study presents rates of closuredue to release from follow-up (or removalat investigator's choice in the CSP).
Dates for each type of termination aredetermined by similar methods in the twostudies. Conception date is estimated astwo weeks after onset of the last men-strual period. For both removal andnoticed expulsion, dates of occurrencewere recorded; unnoticed expulsion isusually considered to have occurred mid-way between the last date the device wasknown or believed to be in place and thefirst date it was missed. Date of loss tofollow-up is recorded as date of lastcontact.
Release from follow-up implies that thedevice is known or believed to be in placeat the time of release. It was dated by theFPS, in contrast to the CSP operationaldefinition, as of clinic cut-off for acceptorswho were overdue at that time and whocame into the clinic later or weirsuccess-fully traced. For those who were notoverdue on the date of clinic cut-off, theFPS date of release from follow-up wasthe date of the first completed follow-upvisit at which the date assigned for the
next scheduled visit fell after clinic cut-offof data collection. In no case did FPS in-clude experience beyond date of last con-tact; whereas CSP included as protecteddays those between last contact and cliniccut-off except in the case of acceptorsoverdue three or more months for ascheduled follow-up.
Unlike the CSP (4), the FPS retained inthe study women reinserted in other thanparticipating clinics (if in the process oftracing we discovered a woman with anapparently continuing segment and thetracing history revealed that there hadbeen an intervening termination and sub-sequent reinsertion) until they terminatedor became lost to follow-up. This practiceseemed reasonable, since study staff hadto trace women overdue by three or moremonths for a scheduled clinic follow-upanyway, and the tracing interview fre-quently was a source of information aboutsuch terminations and subsequent reinser-tions and the months of protection by thenew device. CSP practice was to recordthe preceding termination as an event,record the reinsertion, and close the casein the first month (that is, with one-halfmonth of use).
Comparisons of RatesComparisons of rates obtained in the
two studies are presented in the next fivetables (Tables 3 thru 7). The analyticpopulation in all cases is described asparous acceptors of loop D, ages 15-49,first inserted 31 or more days after lastpregnancy; however, a small number ofnulliparous women (1.8 percent) is in-cluded in the CSP population in Tables3-5. The first three tables present bothcompeting risk (net) and single risk(gross) rates covering two years of ob-servation of the cohort. The differences inthe types of information pmented in eachtable may be discerned in the following:
However, a rough idea of the numbersinvolved may be inferred from the averagenumbers at risk at the bottom of thecolumns. Life table methods of computingthe rates preclude reconstruction of thenumbers of events by simple methods. Inany case the inclusion of confidence inter-vals relieves the reader of the necessity toassess the reliability of rates in the twostudies. For any corresponding pair, theFPS rate is rl.tgarded as significantly differ-ent from the CSP rate if their confidenceintervals are disjoint, that is, if they donot contain common val:res or "overlap".'The FPS rate is underlined in the tableswherever a significant difference occurs.In fact, there are remarkably few sig-nificant differences in the comparablefindings of the two studies.
Under ideal circumstances the CSP andFPS studies would have been jointlyplanned with an eye to comparability.Since they were not, the choices of com-parisons to be made were severely re-stricted by considerations of such com-parability. Thus a number of obviouslyinteresting comparisons that might other-wise have been made, were not.
Interpretation of DataInspection of Tables 3 thru 7 indicates
that, differences in operational proceduresand composition' notwithstanding, thereare no significant differences in pregnancyrates between the two studies for the ob-servation periods indicated, whether onebe speaking of annual or cumulative,competing or single risk rates andwhether one be speaking of first segmentsor all segments for either terminations orclosures.
Virtually the same observation can bemade for first expulsions except for "allsegment terminations" as displayed inTable 4 where the FPS rates are signifi-cantly lower for competing and single risk
Competing risk (net) and single risk (gross) rates
Table 3 first segmentsTable 4 all segments TerminationsTable 5 all segments Closures
Competing risk (net) only
}Annual andcumulative(two-year)
Table AU segments, One-year, controlled for parity within ageTable 7 terminations Two-year cumulative, controlled for parity
Although rate comparisons could havebeen presented for at least four years, thedeterioration in follow-up completenessafter the second year in the FPS elevatedthe hazard of serious bias from thatsource to an unacceptably high level.
Tables of event specific counts havebeen omitted in order to conserve space.
208
A less stringent definition of differencewould have designated two rates as different ifneither fell within the confidence interval of theother. Since confidence intervals are presented,anyone may satisfy himself as to what effectsthis alternative definition would have. Ideallytwo-sample tests using the data from bothstudies would have been used to define significantdifferences, but this would have been time-consuming all out of proportion to its advantage.
TABLE 3. Corriparlsca of CSP and FPS annual and cumulative rates of events per 100 users, by type of termination: loop D,first seginants, two years of observation, porous acceptors ages 15-49 first inserted 31 or mot. days after last preg-nancy.
Competing risk (net) Single risk (gross)
EventsAnnual
Study 1st year 2nd yearCumulative Annual Cumulative(2 years) 1st year 2nd year (2 years)
Pregnancies
Expulsions
RemovalsBleeding/pain CSP 10.9 7.4
(10.2-11.7) ( 6.5- 8.2)
FPS 12.9b 8.8(12.1-13.8) ( 7.9- 9.7)
Other medical CSP 3.1 2.6( 2.6- 3.5) ( 2.1- 3.1)
4.9 4.8( 4.4- 5.5) ( 3.9- 5.3)
0.9 1.9( 0.7- 1.1) ( 1.8- 2.3)
0.7 3.3( 0.5- 1.0) ( 2.7- 3.9)
CSP \ 2.1 2.4C 1.7- 2.4) ( 1.9- 2.9)
CSP 2.4 1.9( 2.0- 2.8)o ( 1.4- 2.3)
FPS 2.5 1.9( 2.1- 2.9) ( 1.4- 2.4)
CSP 9.1 2.3( 8.4- 9.8) (1.8- 2.8)
FPS 8.0 2.1( 7.3- 2.7) ( 1.7- 2.7)
Planningpregnancy
Otherpersonal
Other%
No events
FPS
rCS!?
FPS
FPS r,y 1.4t 1.1
1.6( 1.3- 2.1)
FPS 0.6 1.4( 0.4- 0.9) ( 1.1- 1.9)
CSP 71.5 81.5(70.4-72.6) (80.4-82.8)
FPS 68.9 78.4(67.7-70.1) (74.9-77.7)
3.8( 3.3- 4.2)
3.8( 3.3- 4.3)
10.8
(10.0-11.5)
9.5( 8.8-10.3)
16.2(15.3-17.1)
19.0(18.0-20.0)
4.9( 4.4- 5.5)
8.1( 7.4- 8.8)
2.3( 2.0- 2.6)
3.0( 2.5- 3.5)
3.8( 3.3- 4.3)
2.5( 2.1- 3.0)
1.8( 1.3- 2.0)
58.2(57.1-59.8)
52.6(51.3-53.9)
2.9( 2.5- 3.4)e
3.0( 2.6- 3.5)
9.9( 9.1-10.6)
8.8( 8.0- 9.5)
12.1(11.3-12.9)
14.5b(13.8-15.5)
3.5( 3.1- 4.0)
5.8( 5.2- 6.5)
1.1
( 0.8- 1.4)
0.9( 0.7- 1.3)
2.5( 2.1- 2.9)
1.7( 1.3- 2.1)
0.8( 0.5- 1.1)
2.0( 1.5- 2.5)
2.1
( 1.6- 2.7)
2.5( 1.9- 3.0)
2.4( 1.9- 3.0)
7.8( 6.9- 8.8)
9.4( 8.4-10.4)
2.82.2- 3.4)
5.1
4.4- 6.0)
2.1( 1.8- 2.5)
3.7( 3.0-4.4)
2.6( 2.0- 3.1)
1.9( 1.4- 2.4)
1.7( 1.3- 2.2)
4.9( 4.3- 5.5)
5.0( 4.4- 5.7)
12.1
(11.3-13.0)
11.0(10.1-11.8)
19.0(17.9-20.1)
22.5(21.4-23.7)
6.2( 5.6- 8.9)
10.8( 9.7-11.6)
3.2( 2.9- 3.7)
4.6( 3.9- 5.3)
5.0( 4.4- 5.7)
3.5( 3.0- 4.1)
2.4( 2.0- 3.0)
Average number CSP 5,565 3,202at risk
FPS 3,448 2,178Note: CSP rates from Table 7. page 10. in C. 'natio and S. Leval. "Evalu-
ation of Intrauterine Devices: Ninth Progress Report of the Co-operative Statistical Program." Studies in Family Planning, No. 55(July 1970). Confidence Intervals were calculated by Dr. Tletze's stafffor this comparison.
first expulsion rates in the first annual andtwo-year cumulative categories. With oneminor exception these differences do nottest as significant when simultaneouslycontrolled for age and parity for a one-year period, as presented in Table 6.However, this loss of significance could bedue to the longer confidence intervals oc-casioned by the thinner data when con-trols are introduced.
Later expulsions show significant differ-ences for the second year annual termina-
The perennial dilemma encountered in intro-ducing controls is that one is never sure whetherthe vanishing of significant differences is owed tothe effect of the control or to the diminishedreliability of the rates and the consequentlywider confidence intervals.
4,383
2,812
"Parenthetical figures are the 95 percent confidence limits.'Underlined FPS rates are significantly different from the corresponding
CSP rates.4 Rates of removal for other reasons (believed to be nonrelevant to device
effectiveness) are calculated separately In FPS but are pooled with otherPomona' reasons (fear of injury, etc.) by the CSP.
tion rates in Table 4 but these are not re-flected in the corresponding two-yearcumulative rates.
Removals for pain or bleeding showsignificant differences between the twostudies for first year first segment annualand two-year cumulative terminationrates (Table 3). For all segments however,these differences test as significant only forthe single risk rates (Table 4). For closures,only the two-year cumulative single riskrates show significant differences betweenthe two studies (Table 5). The significanceof differences for first ytar competing riskall-segment terminations vanishes withsimultaneous control of age and parity(Table 6). The same can be said for the
209
two-year cumulative rates when con-trolled for parity except for the parity 1-2category, suggesting that whatever ten-dencies there are to differ are local to thatcategory.
In the category, removals, other medi-cal, the various rates are consistently sig-nificantly higher for the FPS than for theCSP in the absence of controls (Tables 3,4, 5). With age and parity simultaneouslycontrolled, the hypothesis of significantdifferences due to other factors is not sup-ported for competing risk all-segment ter-minations (Table 6). For the two-yearcumulative termination rates for all seg-ments, controlled for parity, significantdifferences occur only in the parity 1-2
. 5
TABLE 4. Comparison of CSP and FPS annual and cumulative rates of events per 100 users, by type of termination: loop D, allsegments, two years of observation, parous acceptors ages 15-49 first inserted 31 or more days after last pregnancy.
Competing risk (net) Single risk (gross)
Events
Pregnancies
ExpulsionsFirst
Annual
Study 1st year
Cumulative ,.,....41.anual Cumulative2nd year (2 years) 1st year 2nd year (2 years)
CSP 2.7( 2.3- 3.1)0
FPS 2.6( 2.2- 3.0)
CSP 9.5( 8.8-10.2)
7.8b
( 7.1- 8.5)FPS
Later CSP 3.2
RemovalsBleeding/pain CSP 11.7
(10.9-12.4)
FPS 12.8(12.0-13.6)
Other medical MP 3.5( 3.0- 3.9)
FPS 4.9( 4.4- 5.4)
Planningpregnancy CSP 0.9
( 0.7- 1.2)FPS 0.7
Other( 0.5- 0.9)
Personal CSP 2.2( 1.9- 2.6)
FPS 1.3( 1.1- 1.7)
Other' FPS 0.6( 0.4- 0.8)
( 2.7- 3.6)
FPS 3.1( 2.7- 3.5)
2.0(1.6-2.4)
2.0(1.5-2.5)
2.5(1.9-3.1)
2.1
(1.7-2.6)
1.7(1.3-2.1)
0.8(0.6-1.2)
7.6(6.8-8.4)
8.9(8.0-9.8)
3.0(2.5-3.5)
4.5(3.9-5.3)
2.0(1.6-2.4)
3.3(2.7-3.9)
2.5(2.0-2.9)
1.6(1.2-2.0)
1.4(1.0-1.8)
4.2( 3.7- 4.7)
3.9( 3.4- 4.4)
11.5(10.7-12.2)
9.2I, 8.5- 9.9)
4.5( 2.0- 5.0)
3.6( 3.2- 4.1)
17.5(16.6 -18.4)
18.7(17.7-19.7)
5.8( 5.2- 6.3)
7.9( 7.2=8.6)
2.5( 2.1- 2.9)
2.9( 2.5- 3.3)
4.2( 3.7- 4.6)
2.4( 2.8)
1.5( 1.2- 1.8)
3.1( 2.6-3.5)0
3.1( 2.7- 3.7)
10.3( 9.6-11.1)
8.6b
( 7.9- 9.4)
3.6( 3.1- 4.1)
3.7( 3.2- 4.2)
12.4(11.6-13.2)
14.6(135.6)
3.9( 3.4- 4.4)
5.9( 5.2=6.5)
1.1
( 0.8- 1.4)
0.9( 0.7- 1.3)
2.6( 2.2- 3.0)
1.6( 1.3- 2.0)
0.7( 0.5- 1.0)
2.1
(1.7- 2.6)
2.2(1.8- 2.8)
2.7(2.2- 3.2)
2.3(1.9- 2.9)
1.0(1.4- 2.2)
0.9(0.6- 1.3)
7.9,(7.1- 8.7)
9.6(8.6-10.6)
3.2(2.8- 3.7)
5.1.(4.4- 5.9)
2.2(1.7- 2.6)
3.7(3.1=4.4)
2.6(2.1- 3.1)
1.8(1.4- 2.3)
1.6(1.2- 2.1)
5.1
( 4.5- 5.7)
5.3( 4.7- 6.0)
12.9(12.1-13.8)
10.1( 9.9-11.6)
5.4( 4.8- 6.0)
4.6( 4.0- 5.2)
19.5(18.4-20.5)
. 22.8(21.6-.24.0)
6.9( 6.3- 7.6)
10.7( 9.8-11.6)
3.3( 2.8- 3.8)
4.6( 3.9- 5.3)
5.1( 4.5- 5.8)
3.4( 2.9- 4.0)
2.3( 1.9- 2.9)
Average number CSP 8,004at risk FPS 3,580
4,127
2,317Note: CSP rates from Table 6, pegs 9, in C. Tietse and S. Lewit, "Evalu-
ation of Intrauterine Devices: Ninth Progress Report of the Co-operative Statistical Program," Stadia. in Family Planning, No. 55(July 1970). Confidence Intervals were calculated by Dr. Tietze's stafffor this comparison.
category for both the competing risk andsingle risk cases (Table 7).
For removals, planning pregnancy wefind significantly higher FPS rates for theuncontrolled cases for second-year annualcompeting and single risk rates and fortwo-year cumulative single risk rates forboth first segment and all segment ter-minations and for all segment closures(Tables 3, 4, 5). Again significant differ-ences vanish for first-year all-segmentcompeting risk terminations when age andparity are simultaneously controlled(Table 6). Significant differences also van-ish for both competing and single risktwo-year cumulative termination rateswith parity controlled. It would seem rea-sonable, in this case at least, to conclude
5,065
2,938° Parenthetical figures are the 95 percent confidence limits.b Underlined FPS rates are significantly different from the corresponding
CSP rates.c Rates of emoval for other reasons (believed to be nonrelevant to device effectiveness) are calculated separately In FPS but are pooled withother personal reasons (fear of Injury, etc.) by the CSP.
that the control varirAes would accountfor the differences observed in the un-controlled cases.
For removals, other personal we findsignificantly low .1. FPS two-year cumula-tive competing risk rates for both firstsegment and all-segment terminations andfor all-segment closures. The same is truefor the single risk termination rates, butthere are no significant differences in thesingle risk closure rates (Tables 3, 4, 5).Again all the differences vanish in thecontrolled cases presented in Tables 6 and7, and it seems reasonable to suppose thatage and parity could account for thedifferences in the uncontrolled cases.
No comparison can be made for re-movals for other reasons tecause these
210
cases were included in other personalreasons in the CSP.
The available information indicateshigher "success" rates for CSP than forFPS. The rates :ar the category no eventsin Table 3 indicate significantly fewer firstsegment terminations for CSP for the firstand second year annual and two-yearcumulative cases. A higher degree of suc-cessful use of the IUD is also indicated forthe CSP population by their significantlyhigher annual and two-year cumulativecontinuation rates. Continuation rates arelower in the FPS than in the CSP, in partat least because of the digerent assump-tions made and procedu aM'followed re-garding experience of the user between thelast follow-up at which the device was
6
TABLE 5. Comparison of CSP and FPS annual and cumulative rates of closures per 100 users, by type of termination: loop D,all segments, two years of observation, parous acceptors ages .15-49 first inserted 31 or more days after last preg-nancy.
Closures
Competing risk (net) Single risk (gross)
Annual Cumulative Annual CumulativeStudy 1st year 2nd year (2 years) 1st year 2nd year (2 years)
Pregnancies CSP 2.4 1.6 3.6 2.7 1.7 4.3( 2.0- 2.7)0 (1.2-1.9) ( 3.1- 4.1)r ( 2.3- 3.1)a (1.3-2.1) ( 3.8- 4.9)
FPS 2.6 1.8 3.9 2.9 2.0 4.8Expulsions ( 2.2- 3.0) (1.4-2.3) ( 3.4- 4.4ryvk ( 2.5- 3.5) (1.5-2.5) ( 4.2- 5.5)
First CSP 2.9 0.9 3.6 3.2 0.9 4.1( 2.5- 3.3) (0.6-1.1) ( 3.1- 4.0' 2.7- 3.6) (0.6-1.2) ( 3.5- 4.6)
FPS 3.4 1.1 4.2 3.7 1.2 4.9( 3.0- 3.9) (0.8-1.5) ( 3.7- 4.8) 3.2- 4.3) (0.9-1.6) ( 4.3- 5.5)
Later CSP 1.9 0.7 2.4 2.2 0.7 2.9( 1.6- 2.2) (0.4-0.9) ( 2.0- 2.8) ( 1.8- 2.5) (0.4-1.0) ( 2.4- 3.3)
FPS 1.7 0.3 1.9 2.0 0.3 2.3Removals ( 1.4- 2.1) (0.1-0.5) ( 1.6- 2.3) ( 1.6- 2.4) (0.2-0.6) ( 1.9- 2.7)
Bleeding/pain CSP 10.4 6.3 15.3 11.0 6.6 16.9( 9.6-11.1) (5.6-7.1) (14.4-16.2) (10.3-11.8) (5.9-7.4) (16.0-17.9)
FPS 11.4 7.5 17.0 12.4 7.9 19.3b(10.6-12.2) (6.7-8.4) (16.0-18.0) (11.5-13.3) (7.1-8.9) (18.3-20.5)
Other medical CSP 2.5 2.2 4.2 2.8 2.4 5.0( 2.1- 2.8) (1.8-2.6) ( 3.7- 4.7) ( 2.3- 3.2) (1.9-2.8) ( 4.4- 5.6)
FPS 4.0 3.7 6.8 445 441 845
Planning ( 3.5- 4.5) (3.2-4.4) ( 6.1- 7.5) ( 4.0- 5.1) (3.5-4.9) ( 7.7- 1.3)pregnancy CSP 0.6 1.5 1.8 0.7 1.7 2.4
( 0.4- 0.8) (1.2-1.9) ( 1.5- 2.2) ( 0.5- 1.0) (1.3-2.0) ( 1.9- 2.8)FPS 0.6 2.6 2.6 0.8 2.9 3.7
Other ( 0.5- 0.9) (2.2-3.2) ( 2.2- 3.1) ( 0.6- 1.1) (2.4-3.5) ( 3.1- 4.3)personal CSP 1.9 2.1 3.5 2.2 2.2 4.4
( 1.6- 2.3) (1.6-2.5) ( 3.1- 4.0) ( 1.8- 2.6) (1.8-2.7) ( 3.8- 4.9)FPS 1.4 1.5 2.6 1.6 1.7 3.3
( 1.2- 1.8) (1.2-2.0) ( 2.2- 3.0) ( 1.3- 2.0) (1.3-2.2) (. 2.8- 3.9)Others FPS 0.5 0.7 1.0 0.5 0.8 1.3
( 0.3- 0.7) (0.4-1.0) ( 0.7- 1.3) ( 0.4- 0.8) (0.5-1.2) ( 1.0- 1.7)Continuation CSP 77.4 84.7 65.6
(76.4-78.4) (83.7-85.9) (64.5-66.8)FPS 74.4 80.7 60.0
(73.3-75.5) (79.4-82.0) (58.7-61.3)
Average number CSP 6,004 4,127 5,065at risk FPS 3,560 2,317 2,938
Noll: CSP rates from Table 5, page 9, in C. nett. and S. Lawn, "Evalu-ation of Intrauterine Devices: Ninth Progress Report of the Co-operative Statistical Program," Studios In Family Planning. No. 55(July 1970). Confidence intervals were calculated by Dr. Tletze sstaff fo r this cson.
° Parenthetical figures are the 95 percent confidence limits.
verified as in place and the cut-off date ofthe study, The CSP includes that experi-ence as continued protection; the FPSexcludes all of it. As M.-Francoise Hallhas indicated, the former method over-estimates the true rate of continuation,whereas the latter method underestimatesit (5).41 FPS one-year all segments con-
In a personal communication Dr. Tietzestates: "In my opinion, the underestimationdescribed here is counterbalanced at leastPartially, by wiping out the experience of allcases lost to follow-up to the last contact. In anyevent, the effect of this difference in procedure issmallest in the first year of use rand increaseswith duration.
.
b Underlined FPS rates are significantly different from the correspondingCSP rates.
c Rates of removal for other reasons (believed to be nonrelevant to de-vice effectiveness) are calculated separately In FPS but are pooled withother personal reasons (fear of Injury, etc.) by the CSP.
tinuation rates are higher than other ratesbased on the same approach of excludingexperience between last visit and cut-off-74 percent compared with 63 percent inthe Baltimore study. The higher parity ofFPS acceptors may account for thatdifference.
All things considered the FPS findingswould appear to justify reinforced confi-dence in the CSP findings. Even wherestatistically significant differences are ob-served, the relative magnitudes of thevarious categories of termination exhibithighly similar patterns. Moreover even
211
without controls the number of cases inwhich no significant rate differences areobserved is quite impressive. Finallywhere the age and parity controls havebeen employed significant differences arequite rare. In Table 6 for instance, whereage and parity are simultaneously con-trolled, only the single subcategory "firstexpulsions, ages 15-24, parity 3-4" showsa significant difference between the twostudies. In Table 7 only parity is con-trolled, but even there, only ten cells outof a possible 48, or about 20 percent showsignificant differences.
..
.
TA
BLE
6. C
ompa
rison
of C
SP
and
FP
S c
ompe
ting
risk
(net
) cu
mul
ativ
e ra
tes
of e
vent
s pe
r 10
0 us
ers
by ty
pe o
f ter
min
atio
n,by
par
ity w
ithin
age
: loo
p D
, all
segm
ents
, one
year
of o
bser
vatio
n.
15-2
4 Y
ears
25-2
9 Y
ears
30-3
4 Y
ears
35-4
9 Y
ears
Typ
e of
term
inat
ion
Pre
gnan
cies
Exp
ulsi
ons
Firs
t
Pari
tyPa
rity
Stud
y 1-
23-
4Pa
rity
Pari
ty5+
1-2
Pari
ty3-
4Pa
rity
5+Pa
rity
Pari
ty1-
23-
45+Pa
rity
Pari
ty1-
2Pa
rity
Pari
ty3-
45+
CSP
FPS
CSP
FP
S
Late
rC
SP
FP
S
Rem
oval
sB
leed
ing/
pain
CS
P
FP
S
CS
P
FP
S
Pla
nnin
gpr
egna
ncy
CS
P
FP
S
CS
P
FP
S
Oth
ers
FP
S
Oth
er m
edic
al
Oth
erpe
rson
al
3.8
3.1
( 2.
8- 4
.8)e
( 1
.8-
4.3)
3.0
2.5
( 2.
3- 3
.9)
( 1.
6- 3
.7)
16.3
13.2
(14.
5-18
.2)
(10.
8-15
.5)
13.0
8.26
(11.
6-14
.7)
( 6.
6-10
.1)
5.9
( 4.
7- 7
.1)
5.1
( 4.
2- 6
.3)
13.9
(12.
2-15
.6)
15.6
(14.
0-17
.3)
4.0
( 3.
0- 5
.0)
4.2
( 3.
4- 5
.3)
2.0
( 1.
3- 2
.7)
1.8
( 1.
3- 2
.5)
2.6
( 1.
8 -3
.5)
1.5
( 1.
0- 2
.2)
0.6
( 0.
1- 1
.1)d
4.6
3.1-
6.2
)3.
32.
3- 4
.7)
9.7
7.6-
11.8
)11
.49.
6-13
.6)
3.9
2.5-
5.3
)5.
64.
3- 7
.2)
0.8
0.3-
1.7
)0.
30.
1- 0
.9)
1.8
0.8-
2.7
)1.
10.
6- 2
.0)
0.5
0.0-
1.1
)d
(2.4
612.
3)(
1.5-
4.0
)2.
11.
6(0
.7-
5.5)
( 0.
5- 4
.2)
10.5
10.5
(5.0
-16.
0)(
8.1-
12.8
)7.
16.
2(4
.0-1
1.8)
( 3.
7- 9
.9)
1.7
2.7
(0.2
- 6.
1)(
1.4-
4.0
)4.
92.
6(2
.5-
9.2)
( 1.
1- 5
.5)
7.5
13.2
(3.4
-14.
1)(1
0.6-
15.8
)10
.518
.5(6
.8-1
5.8)
(14.
2-23
.7)
3.2
3.7
(0.9
- 8.
1)(
2.3-
5.2
)9.
11.
0(5
.6-1
4.2)
( 0.
2- 3
.4)
0.0
1.7
(0.0
- 0.
8)(
0.7-
2.7
)0.
70.
0(0
.0-
3.5)
( 0.
0- 2
.2)
4.3
2.4
(1.4
-10.
0)(
1.2-
3.5
)1.
40.
5(0
.3-
4.6)
( 0.
0- 2
.7)
0.8
(0.0
- 2.
6)d
( 0.
0- 2
.6)
2.7
( 1.
6- 3
.8)
3.3
( 2.
2- 5
.0)
7.5
( 5.
7- 9
.3)
4.3
( 2.
9- 6
.1)
3.1
( 1.
9- 4
.3)
2.9
( 1.
8- 4
.5)
11.8
( 9.
6-14
.0)
13.1
(10.
8-15
.9)
3.3
( 2.
1- 4
.5)
5.4
( 3.
9- 7
.4)
0.6
( 0.
2- 1
.4)
0.2
( 0.
0- 1
.0)
3.1
( 19
-4.
2)1.
0(
0.4-
2.1
)0.
6(
0.2-
1.6
)
3.7
1.4
(1.8
- 5.
6)(0
.4-
3.1)
2.9
(1.7
- 4.
9)5.
16.
1(2
.9-
7.3)
(3.7
- 8.
6)5.
4-
(3.7
- 7.
7)2.
72.
5(1
.1-
4.4)
(1.1
- 4.
6)1.
5-
(0.7
- 2.
9)8.
413
.2(5
.6-1
1.1)
(9.7
-16.
7)8.
9(6
.8-1
1.6)
3.5
2.4
(1.6
- 5.
3)(1
.1-
4.6)
5.1
(3.5
- 7.
3)0.
31.
1(0
.0-
1.4)
(0.3
- 2.
8)0.
2-
(0.0
- 1.
3)2.
51.
3(1
.2-
4.5)
(0.4
- 3.
1)1.
5(0
.7-
2.9)
0.9
0.0
(0.3
- 2.
4)(0
.0-
6.2)
2.2
( 0.
9- 3
.5)
3.4
( 1.
7- 6
.4)
5.3
( 3.
3- 7
.3)
4.3
( 2.
4- 7
.5)
1.0
( 0.
3- 2
.3)
0.5
( 0.
0- 2
.4)
11.6
( 8.
7-14
.4)
14.0
(10.
3-18
.6)
2.0
( 0.
9- 3
.6)
3.4
( 1.
7- 6
.4)
0.0
( 0.
0- 0
.7)
0.5
( 0.
0- 2
.5)
1.2
( 0.
4- 2
.6)
0.0
( 0.
0- 2
.0)
1.4
( 0.
5- 3
.9)
1.4
0.6
(0.5
- 3.
0)(0
.1-
2.1)
2.3
(1.3
- 4.
1)5.
44.
6(3
.2-
7.5)
(2.4
- 6.
9)4.
3(2
.8-
6.5)
2.3
0.0
(1.1
- 4.
3)(0
:0-
1.1)
1.8
(0.9
- 3.
5)8.
613
.6(5
.9-1
1.3)
(9.9
-17.
4)8.
6(6
.4-1
1.3)
4.1
4.2
(2.2
- 6.
0)(2
.0-
6.4)
6.8
(4.9
- 9.
3)0.
50.
3(0
.1-
1.7)
(0.0
- 1.
8)0.
3(0
.0-
1.3)
3.0
1.2
(1.4
- 4.
6)(0
.3-
3.0)
1.8
(0.9
- 3.
4)0.
52.
8(0
.1-
1.7)
(0.2
-13.
1)
0.9
1.9
(0.2
- 2.
5)(0
.8-
3.9)
0.8
0.9
(0.1
- 4.
0)(0
.3-
2.5)
3.6
4.3
(1.7
- 5.
5)(2
.2-
6.3)
4.9
4.9
(2.3
- 9.
5)(3
.2-
7.4)
0.8
0.3
(0.2
- 2.
3)(0
.0-
1.5)
1.6
0.6
(0.3
- 5.
2)(0
.1-
2.0)
10.7
6.8
(7.5
-13.
8)(4
.2-
9.4)
9.5
7.2
(5.8
-15.
1)(5
.1-1
0.0)
2.6
(1.3
- 4.
8)(1
.1-
4.6)
3.1
4.9
(1.2
- 7.
3)(3
.2-
7.4)
0.3
0.0
(0.0
- 1.
4)(0
.0-
1.0)
0.0
0.0
(0.0
- 3.
2)(0
.0-
1.2)
1.6
1.7
(0.6
- 3.
4)(0
.6-
3.7)
3.2
1.5
(1.2
- 7.
4)(0
.7-
3.3)
1.6
0.3
(0.3
- 5.
2)(0
.0-
1.6)
Num
ber
of fi
rst
inse
rtio
nsC
SP
Ave
rage
num
ber
FP
S
at r
isk
CSP
FPS
1,76
4
1,34
5
1,32
696
8
861
746
690
587
128
142
106
113
727
200
566
155
918
532
756
421
Not
e: C
SP
rat
es w
ere
prep
ared
sim
ulta
neou
sly
with
(1]
tabu
latio
ns a
nd s
ent t
o F
PS
for
com
paris
on(L
ette
r of
tran
smitt
al, 1
0/23
/70.
) C
SP
con
fiden
ce In
terv
als
wer
e su
pplie
d by
Dr.
Tie
tze'
s st
aff
for
this
com
paris
on (
6/71
).-
Indi
cate
s ra
te b
ase
(ave
rage
num
ber
at r
isk
In th
e fir
st y
ear
of u
se)
<10
0.P
aren
thet
ical
figu
res
are
the
95 p
erce
nt c
onfid
ence
lim
its.
b U
nder
lined
FP
S r
ates
are
sig
nific
antly
diff
eren
t fro
m th
eco
rres
pond
ing
CS
P r
ates
.C
Rat
es o
f rem
oval
for
othe
r re
ason
s (b
elie
ved
to b
e no
nrel
evan
tto
dev
ice
effe
ctiv
enes
s) a
re c
alcu
late
dse
para
tely
by
FP
S b
ut a
re p
oole
d w
ith o
ther
per
sona
l rea
sons
(fe
ar o
f inj
ury,
etc
.) b
y th
e C
SP
.
L
d
437
431
366
348
393 68 319 54
537
214
449
172
467
387
383
322
378 36 292 28
416
124
350
109
393
330
327
283
The
se th
ree
conf
iden
ce In
terv
als
(the
rat
es fo
r w
hich
are
the
diffe
renc
e be
twee
n to
tal
term
inat
ions
and
the
sum
of p
regn
anci
es, e
xpul
sion
s, a
nd r
emov
als
for
pain
/ble
edin
g, o
ther
med
ical
, pla
nnin
gpr
egna
ncy,
and
oth
er p
erso
nal r
easo
ns, s
ince
we
omitt
ed o
ther
rem
oval
s fr
om th
e sp
ecia
l ana
lysi
sfo
r th
is a
ge g
roup
) w
ere
calc
ulat
ed b
y ha
nd u
sing
the
estim
atio
n fo
rmul
a
SE
-F
rio04
1)10
0T
and
esta
blis
hing
the
95%
upp
er a
nd lo
wer
lim
its o
f the
Inte
rval
as
±2S
E. S
ee T
letz
e(1
) fo
r fu
rthe
rdi
scus
sion
of t
he m
etho
d.
TA9LE 7. Comparison of CSP and FPS cumulative rates of events per 100 users by typo of termination, by parity at first inser-tion: loop D, all segments, two years of observation, parous acceptors ages 15-49 first Inserted 31 or more days afterlast pregnancy.
EventsParity 1-2
CSP
Pregnancies 4.6
Expulsions ( 3.8- 5.4)aFirst 14.3
(13.0-15.6)
Later 5.7
Removals ( 4.8- 6.5)Pain /bleeding 19.4
(18.0-20.9)
Other medical 5.74.8- 6.6)
Planning pregnancy 4.7( 3.9- 5.5)
Other personal 4.4( 3.6- 5.2)
Other
Pregnancy 5.8( 4.8- 6.8)a
ExpulsionsFirst 16.4
(14.9-17.9)
Later 7.1
Removals( 8.0- 8.2)
Pain/bleeding 22.1(20.4-23.7)
Other medical 7.1( 6.0- 8.2)
Planning pregnancy 8.5( 5.4-, 7.6)
Other personal 5.8( 4.8- 6.9)
Other
Average numberat risk 2,052
FPS
Parity 3-4
CSP FPSParity 5+
CSP FPSCompeting risk (net) rates
3.8( 3.1- 4.7)
13.6(12.2-15.1)
5.2( 4.3- 6.2)
23.1(21.4-24.9)
6.4( 5.4- 7.5)
5.9( 4.9- 7.0)
2.5( 1.9- 3.3)
1.4( 0.9- 2.0)
4.1( 32- 4.9)
10.2( 9.0-11.4)
4.3( 3.4- 5.1)
17.0(15.5-18.5)
5.9( 4.9- 6.9)
1.1( 0.7- 1.6)
3.9( 3.1- 4.7)
3.8 3.8 4.0( 3.0- 4.7) ( 2.7- 4.9) ( 3.1- 5.1)
7.3a 7.2 5.8( 6.3- 8.5) ( 5.8- 8.6) ( 4.8- 7.0)
3.1 2.4 2.22.5- 4.0) ( 1.5- 3.2) ( 1.5- 3.0)
18.4 13.6. 13.1(16.8 -20.1) (12.8-14.3) (11.6-14.9)
8.3 5.5 9.3( 7.2=9.8) ( 4.2- 6.8) ( 8.0-10.8)
1.4 0.3 0.8( 0.9- 2.0) ( 0.1- 0.8) ( 0.4- 1.4)
2.1 4.3 2.5( 1.6- 2.9) ( 3.1- 5.4) ( 1.9- 3.4)
1.6 1.5( 1.1- 2.3) ( 0.9- 2.2)
Single risk (gross) rates
5.7( 4.8- 7.0)
16.7(15.0-18.5)
7.3( 6.1- 8.7)
30.0b(27.9-32.3)
10.0( 8.6-11.8)
10.9( 9.2-12.8)
4.0( 3.1- 5.3)
2.6( 1.8- 3.8)
4.8( 3.9- 5.8)
11.2( 9.9 -12.6)
5.1( 4.1.; 13.1)
18.6(17.0-20.3)
7.0( 5.8- 8.1)
1.4( 0.9- 2.0)
4.7( 3.7- 5.6)
4.9 4.3 5.2( 4.0- 6.1) ( 3.1- 5.6) ( 4.1- 6.6)
8.5 7.9 6.5( 7.3- 9.9) ( 6.4 -9.5) ( 5.4- 7.9)
3.8 2.7 2.6( 3.0- 4.8) ( 1.7- 3.7) ( 1.9- 3.6)
22.0 14.7 15.4(20.1-24.0) (12.8-16.8) (13.6-17.3)
10.9 6.2 11.2( 9.4-12.5) ( 4.8- 7.7) ( 9.7-13.0)
2.1 0.4 1.1( 1.4- 3.0) ( 0.1- 1.0) ( 0.6- 1.9)
3.0 4.9 3.2( 2.2- 4.0) ( 3.6- 6.2) ( 2.4- 4.4)
2.4 2.0( 1.7- 3.4) (1.3-3.0)
954Not.: CSP rates from C. Tian and S. Lewit. "Evaluation of Intrauterine
Devices: Ninth Progress Report of the Cooperative Statistical Pro-gram. studies In Family Manning. No. 55 (July 1970). CSP 95%confidence intervals were calculated by Dr. Tiefte's staff for thiscomparison (8/71).
It can, then, be concluded that a reason-ably approximate replication of the CSPstudy has confirmed its general validity.
References1. Christopher Tietze and Sarah Lewit.
"Evaluation of intrauterine devices: Ninthprogress report of the Cooperative Statis-tical Program," Studies in Family Planning1, no. 55 (July 1910).
2. Christopher Tietze. "Complications with
1,915 1,068 1,038 916Parenthetical figures are the 95 percent confidence Intervals.
b Underlined FPS rates are significantly different from the correspondingCSP rates.
IUD's." British Medical Journal 2 (5654)(17 May 1969): 444-5.
3. Roger R. Bernard, M.D. "I.U.D. Perform-ance Patterns-A 1970 World Viet- ."International Journal, of Gynecolay andObstetrics (November 1970), Vol. 8, No. 6,Part 2, p. 934. Clinic differences in rates ofcontinuation and of removal for painand/or bleeding persist only within the setof clinics having at least 10 percent of firstacceptors of loops B, C, or D lost tofollow-up within the first year and first
213
segment, when clinic differences in dis-tribution of first insertions by loop sizehave been taken into account.
4. Letter from Christopher Tietze, M.D.,10/23/70.
S. M.-Francoise Hall. "Field effectiveness ofthe oral and the intrauterine methods ofcontraception: The Baltimore Public Pro-gram, 1964-1966." Milbank MemorialFund Quarterly 47, no. 1, Part 1 (January1969): 65-66.
Policies Affecting Population in West Africa
by DAVIDSON R. GWATKIN
Mr. Gatkin is program advisor in population in the West Africa Office of theFord Foundation, in Lagos, Nigeria. "West Africa" as used in this article refersto the 20 countries served by the Ford Foundation's regional office in Lagos:Cameroon, Central African Republic, Chad, Congo (Brazzaville), Dahomey',Gabon, The Gambia, Ghana, Guinea, Ivory Coast, Liberia, Mali, Mauritania,Niger, Nigeria, Senegal, Sierra Leone, Togo, Upper Volta, and Zdire (formerlyCongo / Kinshasa). The article is based on a paper to be published in a forthcomingbook, Population Growth and Socio-Economic Change in West Africa, editedby John Caldwell, Nelson Addo, S. K. Gaisie, Adenola 'gun, and Philip Olusanya.
In West Africa, the unintended conse-quences of general government develop-ment programs are far more important ininfluencing population growth and move-ment than are explicit government popu-lation policies. Very few of the region's20 countries have expressed official con-cern about population, but all of them areactively promoting social and economicdevelopment. The overall impact of gov-ernmental development policies has beento increase greatly the rate of populationgrowth and the movement of people fromrural to urban areas. Governments thatsay or do nothing explicitly about popula-tion are, thus, far from neutral. Withoutrealizing it they are the active promotersof increasing population growth andrural-urban migration.
GOVERNMENT POPULATIONPOLICIES
Policies on GrowthWent African governments fall into
three categories with regard to populationpolicy: those concerned about populationgrowth, those that have expressed little orno awareness of the issue, and those eagerto have their populations grow morerapidly. The second of these three groupsis by far the largest, constituting roughlythree-quarters of the governments in WestAfrica.
FOR LESS RAPID GROWTH:GHANA AND NIGERIA
The West African government mostconcerned with reducing its rate of popu-lation growth is Ghana. In March 1969Ghana became the first country in WestAfrica to define an official populationpolicy. In its policy paper (Ghana, 1969),the government expresses its concern withreducing growth:
The population of Ghana is the nation'smost valuable resource. . . . The protection
and enhancement of its welfare is the Govern-ment's first responsibility. When that welfareis threatened, the Government must act. Thewelfare of the nation is now endangered by asubtle, almost imperceptible demographicchange. . . . Unless birthrates can be broughtdown to parallel falling death rates, Ghana'spopulation will climb at a rate dangerous tocontinuing prosperity, and the children of thenext few generations will be born into a worldwhere their very numbers may condemn themto life-long poverty (p. 3).
The policy statement begins with adescription of Ghana's demographic situ-ation and the causes of rapid populationgrowth in the country and then examinesthe consequences of this growth for in-come, food supply, educational expan-sion, employment, health, and welfare.There follows a discussion of the feasi-bility of lowering population growththrough use of modern contraceptivemethods. One of the seven specific policiespresented in the statement declares:
Recognizing the crucial importance of awide understanding of the deleterious effectsof unlimited population growth and of themeans by which couples can safely and effec-tively control their fertility, the Governmentwill encourage and itself undertake programsto provide information, advice, and assistancefor couples wishing to space or, limit theirreproduction. These programs will be educa-tional and persuasive, and not coercive (p. 20).
This proposal resulted in the establish-ment of an active, adequately-fundednationwide family planning program in1970.
The statement also proposes steps be-yond family planning:
Ways will be sought to encourage and pro-mote wider productive and gainful employ-ment for women; to increase the proportionof girls entering and completing school; todevelop a wider range of non-domestic rolesfor women; and to examine the structure ofGovernment prerequisites and benefits and ifnecessary change them in such ways as tominimize their pronaialist influences andmaximize their antinatalist effects (p. 20).
214
This provision; 's especially applauded bythose who we that family planningprograms alone are not sufficient to bringabout a significant reduction in fertilityand that strong social measures are alsoneeded. However, iodate, the governmenthas not been as active in implementingthis element of its polizy as it has been ininitiating the family planning program.
Although official concern with popula-thIn growth was expressed much earlier inNigeria th;:tn in Ghana, it remains moretentative thcre. Reference to the need fora copulation policy appeared as early asJune 1966 in Guideposts for the SecondNctional Development Plan, prepared bythe Ministry of Economic Development.Title VII of that document, entitled "Pop-ulation growth," tentatively placed thegrowth rate in Nigeria between 2.0 and
.5 percent per year. It called for a demo-graphic sample survey to provide moreaccurate information and declared, "Thereis a need to evolve a population policy(including spatial distribution) as anintegral part of the national developmenteffort" (Nigeria, 1966. ,. 3).
Since then, increasing numbers of gov-ernment leaders in Nigeria have publiclyexpressed concern about populationgrowth and the need for family planning.Those issuing statements have includedthe Federal Government Commissionersof five ministries (those of Economic De-velopment and Reconstruction, Agricul-ture and Natural Resources, Health, In-formation and Labour, and Education);two Federal Government PermanentSecretaries (those of Health, and of Eco-nomic Development and Reconstruction);the Military Governors of three out of 12states (those of Lagos, Kano, and North-Central); and the Commissioners ofHealth and Social Welfare of five states(those of Lagos, Western, Mid-Western,Kwara, and North-Eastern) (Annual Re-ports, 1967, 1968, 1969). Although theirstatements express suppgrt for familyplanning, citing the addition i :.enefits ofmaternal and child health .7:ire, familywelfare, and the human right to determinethe number of pregnancies, not all makespecific reference to a possible relationbetween family planning and populationgrowth.
State and local governments in Nigeriaprovide support for family planningforexample, through their cooperation withthe private Family Planning Council ofNigeria. Several of these governmentshave allowed the council to operate familyplanning clinics in government facilities
10
whenever these are not otherwise in use.Many also a:low gover-me,:t health per-sonnel to work in the Family PlanningCouncil clinics outside their regularhours, to participate in training courses,and, in some cases, to work :1111-time forthe council.
In November 1970, the Nigerian FederalGovernment issued its official populationpolicy as part of the second NationalDevelopment Plan (1970-1974). The state-ment (Nigeria, 1970) begins by stressingthe potentially controversial nature of apopulation policy in Nigeria. It then out-lines briefly the country's demographicsituation, notes the problems associatedwith rapid population growth, but alsopoints to the Nigerian resources and de-velopment potential as factors offsettingthese problems. It concludes:
The magnitude of the country's populationproblem is unlikely to be such that calls foremergency or panic action. . . . What seemsappropriate in the present circumstances ofNigeria is for the Government to encouragethe citizens to develop a balanced view of theopportunities for individual family planningon a voluntary basis, with a view to raisingthe quality of life in their offsprings. Facilitiesare to be designed to protect mothers, on along-range basis, from repeated and un-wanted pregnancies, as well as to enableparents to space their children for betterfeeding, clothing, and education . . . duringthe Plan period, the Government will pursuea qualitative population policy by integratingthe various voluntary family planning schemesinto the overall health and social welfare pro-gram of the country. . . . The Governmentwill establish a National Population Councilto implement this population policy and pro-gram, and to coordinate all external aid sup-port for family planning activities throughoutthe country (pp. 77-78).
Thus Nigeria has committed itself tothe provision of family planning servicesthrough existing health facilities. But theNigerian policy statement clearly placesmuch less emphasis on the need for re-ducing the rate of population growth thandoes that of Ghana.
THE UNCONCERNED WEST AFRICANMAJORITY
Outside of Ghana and Nigeria, officialconcern about poptes:ion growth islimited. In most countries a few indi-viduals from the health fields are at-tempting to alert their governments to thevalne of family planning. At the sametime, in many o 'hese countries policymakersparticu tat iy those assoc iated witheconomic planningare becoming awareof the lack of knowledge about popula-tion growth in their countries and are
seeking ways to obtain further informa-tion. But concerned individuals are stillvery mu':h in the minority.
The prevailing attitudes can be bestunderstood by briefly considering a fewcountries that are representative of the15* in this category:
Dahomy. Daho:; economic plan-ners displayed wi es probably theearliest official cot, .t,t in West Africawith population growth in the 1961-1965Economic Development Plan (Dahomey,1961).
The Dahomean population is growingrapidly, at 2.65 percent per year. . . . It isdoubling every 26 years. Half the population,is massed in the South, in a coastal band rep-resenting 7 percent of the country's land area,which is heading toward a high density. TheDepartments of the South-East and theNorth-West find themselves grappling withformidable problems of local over-popula-tion. Forty-six percent of the population isless than 15 years of age, which renders moreacute the problems of food, education, andemployment (p. 175).
Altho,;gh this analysis was not fol-lowed by action, individuals within thegovernment remained concerned. Thisconcern led to a government request thatthe Population Council send a mission tostudy the country's population situationin June 1969. The final mission report wassubmitted in the summer of 1970, but thegovernment's reaction is not yet known.To date, the country has no populationpolicy.
The Gambia. Interest in the size andgrowth of Gambia's population wasspurred by the 1965 publication of thefinal 1963 census results. By 1968 medicalinterest in family planning had increasedenough that a private family planningassociation was established with moralsupport from the government. Subse-quent reports on The Gambia economyby the International Monetary Fund andby a British Ministry of Overseas De-velopment expert stressed the potentiallynegative effect of population growth onthe country's economic growth.
As a result of this concern, the govern-ment requested a Population Councilmission to study the available data andrecommea an appropriate governmentpolicy. This study was undertaken in thesummer of 1969. Its final recommenda-
That is, all countries not specifically namedas favoring less or more rapid populationgrowth. These unconcerned countries arc:Central African Republic, Chad, Congo (Brazza-ville), Dahomey, The Gambia, Guinea, Liberia,Mali, Mauritania, Niger, Senegal, Sierra Leone,Togo, Upper Volta, Zaire.
215
tion: ..,ere submitted to the government inJune 1970, but as no decision on themhas been taken.
Senegal. Several leuders have becomeconcerned over the impact of rapid popu-lation growth and have planned a specialfertility rate survey (or 1970-1971 to ob-tain the necessary information for furtherconsideration of the issue. In the mean-time, the government has given informalbut strong encouragement to the pioneer-ing founders of the first family planningassociation in French-speaking Africa.
Sierra Leone. The government has re-mained neutral on the question of familyplanning. As Thomas Dow (1969) of thePopulation Council reports:
While the government mainwirsc.-.n officialposition of . . . neutrality with regard tofamily planning, it allows the Planned Parent-hood Association to provide services in publicas well as private facilities.. . . The effect ofexiting planting activities is negli-gible.. Whether this will continue be thecos: depend largely on the governmeorspolicy. If it maintains its present position ofneutrality, this will effectively preclude anysignificant increase in family planning inSierra Leone. Conversely, if it endorses andfinancially supports family planning, sub-stantial progress might be possible in eerie nearfuture. At the moment, there is no indicationthat such a change is in the offing (p..4).
However, this neutrality is a new de-veloomPnt for Sierra Leone. When Joht.,atzt:.7,;..! of Population Council con-ducted e. 1965 survey of governmentalpopulation policy in CommonwealthAfrica, Sierra Leone was distinctly pro-natalist. It was the only country in WestAfrica expressing the belief that it wouldbenefit from a much larger populationand conveying the hope that the birthrate might rise (Caldwell, 1968a, pp. 369,371).
Togo. There appears to be no officialpolicy about population growth and therelated question of family planning inTogo. Present indications are that theTogolese Council of Ministers has de-cided to remain neutral on the subject.Yet this should not be taken to mean thatthe government favors rapid populationgrowth. David and Jeane Stillman (1969)of the American Friends Service Com-mittee, who recently interviewed Togolesegovernment officials, report that:
Those members of the government inter-viewed did not appear to believe that thecountry would benefit from a populationmuch larger than at present, but were con-cerned rather with redistribution and im-proved living standards for those who alreadymake up the populace. The governmentavoids the notion of seeking to fill up empty
lands with a population boom.... Althoughthe rate of population growth is not presentlya matter of alarm, government planners dofeel that rapid growth could hinder attemptsto raise living standards. Those thinking interms of the educational realm are particu-larly ccsnizant of this (p. 8).
FOR MORE RAPID GROWTH
Cameroon. At the 1969 World HealthAssembly in Boston, the Cameroon gov-ernment representative spoke of thealarming population decrease in CentralAfrica because of high mortality, es-pecially among infants. He expressed akeen interest in measures to combatsterility as a means of counteracting this.He also felt that an increase in populationwould aid the developing countries toreach their goals more quickly. More re-cently, other Cameroonian governmentspokesmen have said that no measures toslow growth of the present 6 millionpopulation are to be conts mplated untila goal of 15 million is achieved. The gov-ernment has adopted a variety of policieswith rronatalist aspects for workingwomen, including paid maternity leave,children's allowances; and time off duringoffice hours for care of children. At thesame time, the Cameroon governmentdoes not oppose the use of contraceptivemethods when prescribed by a physicianfor the sake of the mother's health.
Gabon: Gabon has been concerned withincreasing its birth rate for some time.The 1963-1965 Gabon Intermediate De-velopment Plan predicted that the expan-sion of public health programs wouldpromote an increase in the country'spopulation growth rate, which was con-sidered insufficient at that time (de Buffet,1970). The 1966-1971 Gabon Develop-ment Plan predicted a shortage of man-power by the bejnning of 1970 (WestAfrica, 1970, p. 925). According to aUnited Nations Economic Commissionfor Africa survey; "Gabon's aim is tomaintain its present birth rate . . . at 35per 1,000" (United Nations, 1969, p. 12).The government has recently passed a lawthat prohibits importation of contracep-tives, has instituted severe fines and prisonpenalties for any violation of the law, andhas called for the appointment of specialinspectors for its enforcement (Prader-vand, 1970, p. 17).
Such strong measures must be viewedin the context of Gabon's unique geo-graphic and economic situation. Gabonhas by far the highest per capita incomein West Africa, estimated at approxi-mately USS400 per year in 1966. Its birth
rate is 'oy far the lowest in the region.Coupled with a continuing high mor-tality rate, this birth rate produces anextraordinarily low population growthrate of about 1 percent per year. Gabon'saverage population density (around 2.3per square kilometer) is also extremelylow. The country's rich forest and mineralresources require a large labor force fortheir exploitation ("Gabon," 1970, pp.349-352).
Ivory Coast. Recent visitors to the IvoryCoast report a similar but as yet formallyunexpressed governmental interest in pro-moting population growth for economicreasons. The Ivory Coast's gross nationalproduct has been expanding steadily at arate of 7 percent per year. This growthmakes it the only country in West Africaable to finance a rapidly rising standardof living despite a population growth rateestimated at 2.3 percent. However, thiseconomic boom is also attracting a largenumber of migrants from neighboringcountries, and it is now estimated that asmuch as one-quarter of the population isof foreign origin. This has generated con-cern among government authorities aboutthe inability of the labor force to con-tinue to absorb immigrants and about thepolitical problems a large foreign popula-tion might produce.
Policies on MigrationAmong the nations of West Africa, the
issue of migration has received somewhatmore governmental attention than popu-lation growth. Concern has focused upontwo types of migrationmigration acrossinternational boundaries and internalrural-urban migration.
INTERNATIONAL MIGRATION
As John Caldwell reported in his surveyof population policies of CommonwealthAfrica (1968a):
In most countries, the oldest type of popu-lation policy has been that relating to theregulation of migration. Such regulation hashad to be considered more in recent years,because many of what arc now internationalborders were little more than internal bound-aries in the extensive British and French em-pires. Furthermore, independent govern-ments are much more likely to feel a priorresponsibility for securing employment fortheir own nationals than a colonial govern-ment (p. 370).
This attitude is apparent in the presentpolicy of the Ghana government, which,once again, has the most clearly articu-lated position of any West African coun-try. The March 1969 population policystatement declared:
The Government will adopt policies andestablish programmes to . . . reduce the scaleand rate of immigration in the interest ofnational welfare . . . [because] uncontrolledimmigration of labor, especially of the un-skilled type, reduces employment opportuni-ties for citizens. It is intended that immigra-tion will be used primarily as a means ofobtaining needed skills and stimulating socialand economic development (p. 23).
On 18 November 1969, the governmentdecreed that all aliens who did not obtainthe required residence permit within twoweeks would be expelled from the coun-try. No complete official explanation forthis action was given. Various unofficialexplanations were: the desire to createmore jobs for the increasing numbers ofunemployed Ghanaians in the cities, con-cern over high crime rates reported amongalien communities, a political tactic togain prestige by appealing to latent hos-tility of citizens to outsiders, a desire tocurb specific undesirable alien acevitiessuch as domination of the retail sector bySyrians and Lebanese. Whatever the pur-pose, the result was a mad scramble byaliens to obtain residence permits and adramatic exodus of thousands back totheir native countries (particularly UpperVolta, Togo, Dahomey, and Nigeria)(West Africa, 1969 and conversations inAccra). Although there is no way to con-firm the statistics, in January 1970 thegovernment announced that approxi-mately 200,000 of Ghana's 1,000,000alien residents had left (Daily Graphic,1970, p. 1). The continuing efforts by thepolice to apprehend noncomplying aliensleave no doubt about the government'sdetermination to control immigration.
In other countries, governmental con-cern is more sporadic. Among the morenoteworthy expressions of it have been:
Expulsion of Dahomeans and Togo-lese from the Ivory Coast in 1958. AsAkin Mabogunje of Ibadan Universityhas reported, Dahomey supplied manyareas of West Africa with professional aswell as skilled manpower. As the move-ment toward independence gathered mo-mentum, resentment grew. When a newinflux of Dahomean and Togolese immi-grants was rumored in October 1958,violence broke out in Abidjan, capital ofthe Ivory Coast. One man was killed, 50injured, and much property damaged. Asa result, more than 1,000 Togolese andDahomean men and women left the IvoryCoast to return to their homes (Mabo-gunje, n.d., pp. 146-147).
216 12
tr
Expulsion of Dahomeans from Nigerin 1963-1964. Again, Mabogunje reports". . . in 1963, over a dispute with Da-homey, Niger suddenly ordezed all Da-homtans living in its territory, estimatedat about 16,000, to leave the countrybefore the end of the year. Later, in 1964,it agreed to expel only Dahomean civilservants and not all Dahomeans, as hadbeen threatened...." (p. 147).
Expulsion of Ghanaians from SierraLeone. In December 1968, the SierraLeone government ordered 2,000 Gha-naian flrhermen to leave the country be-cause they prevented Sierra Leoneansfrom fishing. The government also chargedthat the Ghanaians did not pay taxes andv. ere a security problem in the event ofattempted sea invasion (pp. 149-150).
The journal West Africa (1969) citesthese and other cases in arguing that theNovember 1968 Ghanaian action wasrelatively mild:
Ghana is not expelling all aliens but onlythose without residence permits; permitswill, no doubt, be readily issued to those incertain occupations such as mining. . . . Thetrend is clear. To a greater or lesser degree, itis the ambition of each of the governments. . . into which West Africa is divided, toreserve the right to live and work within itscountry for those who can pass a restrictivecitizenship test.
Mabogunje has aptly summarized thesituation: "Pan-Africanism appearseverywhere to be on the retreat. It isbeing replaced by economic chauvinismwhich sees the contribution of the mi-grants not as resulting from their specialcircumstances but from an usurpation ofthe 'rights and privileges' of the membersof the host community" (p. 14).
RURAL-URBAN MIGRATION
Many West African governments haveadopted policies to curb rural-urban mi-gration. But these policies have yet to beproven effective. Once more Ghana hasexpressed its views most clearly. TheMarch 1969 government policy statementsays:
The Government will adopt policies andestablish programmes to guide and regulatethe flow of internal migration, influencespatial distribution in the interest of develop-ment progress... . The Government intendsto influence the flow of migration within thecountry so as to avoid excessive and un-controlled concentration of population inurban areas. . . . It intends also to provideincentives for the siting of industries awayfrom the major metropolitan areas; to urgeindustries in both private and public sectors
i
to decentralize some of their functions; tointensify its efforts to supply rural areas withfacilities such as safe pipe-borne water andimproved means of refuse disposal" (p. 22).
Programs to curb rural-urban popula-tion flows have been developed by govern-ments in French-speaking West Africa.
In Zaire (formerly Congo, Kinshasa),an effort was made in 1967 to implementa "return to the land" policy through theestablishment of a youth organization todevelop agricultural communities. Majoremphasis has also been placed on theestablishment of new rural communities.
Senegalese efforts to stem rural-urbanmigration have focused on resettlement.The government seeks to open "newlands" in sparsely settled areas throughthe organization of communities sup-ported Ly adequate infrastructure andcapital investment. This "regional pro-gram" was described as a principal long-term objective of the Second Develop-ment Plan (1965-1969), designed in part"to prepare over the long run a moresatisfactory regional equilibrium" and "toface the demands of demographic growthby increasing present production. . . ."(Senegal, n.d., p. 54).
Other West African governments havedeveloped their own variations of the"return to the land" or "new lands"programs. The Upper Volta Plan Outlineof 1967-1970 emphasized rural develop-ment as a way to stem potential rural-urban migration (Haute-Volta, n.d.). TheNiger 1965-1968 Four-Year Plan ex-pressed its concern at the rapid rate ofurbanization, citing the expense to thegovernment of providing for city-dwellerswhat rural people either go without orprovide for themselves through collectiveaction (Niger, n.d., p. 168). Internal popu-lation distribution and rural-urban popu-lation movements constitute the TogoGovernment's principal population inter-est (Stillman, 1969, p. 5).
Unfortunately, these policies are seldomimplemented, and when they are, theyrarely achieve the desired population re-distribution or increase agricultural pro-duction. As the Congolese agriculturalplanners warned in correctly predictingthe failure of the government's 1967-1968"return to the land" program, urban con-sumers moved to rural areas tend to be-come rural consumers rather than pro-ducers (Congo, Kinshasa, 1966, p. 19).
In references to publications from this area,we have kept the name of the country that ap-pears on the publication.
217
GOVERNMENTAL DEVELOPMENTPOLICIES AND THEIR UNINTENDED
POPULATION EFFECTSAlthough only a few West African gov-
ernments have dealt directly with popula-tion issues, all of them have unintention-ally influenced population growth andmovement through policies directed to-ward other aspects of social and eco-nomic development. The most importantof these policies are in the areas of health,education, family allowance, industriallocation, and government revenue andexpenditure allocation patterns betweenrural and urban areas.
HealthAll West African governments are ac-
tively and effectively seeking to improvetheir citizens' health by establishing andsupporting health programs. Table 1 givesavailable information on health expendi-tures of West African countries. No coun-try listed devotes less than 7 percent of itsrecurrent budget to health. The average isclose to 10 percent. The per-capita healthexpenditure ranges from USS0.68 to S4.40with the average $1.25-51.50. Many ofthese countries also provide additionalcapital investment funds for special healthprograms not included in the figures inTable 1.
These expenditures have produced re-sults. The region-wide smallpox- measlesprogram, for example, has given over 100million smallpox vaccinations with the re-sult that West and Central Africa re-corded only 10 percent as many smallpoxcases in 1969 as in 1968 (Morning Post,1970, p. 12). The Government of Ghanaestimates that life expectancy at birth hasrisen from 28 years in 1921 to 48 in 1965a rate of nearly one-half year per yearover the past 20 years. (Ghana, 1969, p. 6).The mortality level in Nigeria is estimatedto have fallen by one-sixth from 29 per1,000 to 24 per 1,000 in the ten years sinceindependence. Projections of the Nigerianpopulation made by the DemographicUnit of the University of Ife envision acontinuing decrease in the mortality levelto 14 per 1,000 population in 1990 (Cald-well and Igun, 1971). This informationconfirms the prevailing view at the 1966Ibadan Population Conference, as re-ported by John Caldwell: "Despite theuncertainty about levels and trends, theconviction is widespread that death ratesare falling, perhaps rapidly" (Caldwell,1968b, p. 11).
Yet there has been no correspondingdrop in birth rates. This is the basic cause
TABLE 1. Health expenditures of West African governments.
Country Year
A mount ofrecurringbudget
Approximatepercent oftotalgovernmentrecurringbudget Per capita
Dahomey 1965 US$ 3,946,000 12 C.75Ivory CoastNigerSenegal
196519851985-1966
$12,435,000$ 2,258,000$12,232,000
1099
$3.40$0.68$3.65
Upper VoltaMali
19651985
$ 3,335,0004,508,000
1011
$0.70$0.98
CentralAfricanRepublic 1985 $ 2,785,000 10 $1.85
Chad 1965 $ 3,327,000 10 $1.00Congo
(Brazzaville) 1965 $ 3,077,000 8 $3.60Gabon 1965 $ 2,777,000 9 $4.40Cameroon 1965-1966 $ 8,817,000 11 $1.75Togo 1985 $ 1,850,000 9 $1.15Nigeria 1969 $67,200,000 10 $1.02Liberia 1964 $ 2,618,000 7 $2.50
Sources: U.S. National Academy of Sciences. Nativist Research Council, Division of Medical Sci-ences, Public Health Problems in 14 French-speaking Countries in Africa and Madagascar.v.11. A Survey of Resources end Needs. passim.United Nations, World Health Organization. Third Report on the World Health Situation.1961-1964 (Official Records of the World Health Organization, No. 15.5. Gwen% World HealthOrganization. April, 1967. passim.
of the unusually rapid and increasingpopulation growth in West Africa. Thus,governmental disease control programshave beenand will continue to beanimportant factor in accelerating popula-tion growth.
EducationWest African governments have shown
even more interest in education than inhealth. None of the countries in Table 1for which information is available spendsless on education than on health. In mostof them, educational expenditures runtwo, three, or four times higher thanhealth expenditures. As a result ever-increasing numbers of school-age childrenare attending school. In the French-speaking countries of West Africa, for ex-ample, the number of children in schoolincreased by over one-third in the three-year period 1961-1964 (Hallak andPoignant, 1966, p. 18); in Nigeria primaryschool enrollments tripled between 1952and 1960, and secondary school enroll-ments tripled between 1956 and 1966(Nigeria, 1966).
Demographic effects of these govern-ment expenditures on education and re.suiting rises in school enrollments are notcomoletely clear. However there are indi-tations that educational expansion is be-ginning to exert a significant downwardinfluence on birth rates because of twophenomena:
I. Educated couples tend to limit ther umber of their children. John Caldwell
(1968b) summarizes a recent report on thefindings of Sub-Sahara African surveys ofdesired family size by noting the ubiqui-tous importance of education:
. .. in every country where data were pre-sented on the point, the existence and lengthof formal education has been a major de-terminant of behaviour patterns. It influencesattitudes towards family size, knowledge ofcontraception, and practice of it. In Lagos,and the pattern is probably similar else-where, ten times as many women with terti-ary education were practicing birth control aswere uneducated women (p. 618).
2. Couples are under economic pres-sure to limit family size in order to provideeducation for children. Caldwell (1968b)reports that African surveys show that:
the chief problem of the large familyis clearly that it is an economic burden... .Equally clear is the fact that educationalcosts climb steadily from subsistence to cashfarming areas and from the latter to thetowns, culminating among the urban elitewhere these costs are the major factor whichcould lead to the limitation of family size(P. 604).
The impact of educational costs onbirth rates is likely to depend on the levelof those costs. If education were not avail-able at all or available only at exorbitantcost, parents would realize their inabilityto provide education regardless of howseverely they limited their family size andso could be expected to continue theirformer reproductive habits. Inexpensiveor totally free education could also exertno appreciable effect on family size sinceparents would be assured of their chil-
218
dren's education regardless of the numberof children. The incentive for fertilitylimitation comes at a moderate price levelat which pare: can afford education butonly for a limited number of children andonly at considerable sacrifice.
Not enough information exists to proveconclusively that the cost of education toWest African parents is "moderate" orthat educational costs actually influencetheir reproductive behavior. But it can beshown that governments carry by far thegreatest part of educational costs andthat, despite this, parents must still con-tinue to contribute a large amount ofmoney relative to their incomes to meettheir share of expenses.
The government's contribution is welldocumented. The Ghanaian governmentfor example, pays the entire USS20 annualcost for a child's primary school attend-ance. The parents contribute only a $1.50book fee. In secondary schools, the gov-ernment pays about $160 per year and re-quires parents to contribute approxi-mately $11 for books and uniforms(Ghana Ministry of Education, n.d.). InNorthern Nigeria, the government con-tributes approximately $12 a year for eachchild in primary school which covers allexpenses except uniforms (Nigeria FederalMinistry of Education, n.d.). In French-speaking Africa, governments contributeannually between $25 and $75 per child atthe primary school level with parents notrequired to add much more. According toHallak and Poignant's (1966) study ofthis group:
In the aggregate, the [financial] role (ofparents) is very modest: in so far as publiceducation is free, the expenses of familiesrepresent essentially the school fees in theprivate institutions of instruction. However,in certain cases, the initiative of individualstakes the form of "in kind" financing, varyinggreatly with the particular conditions of theregion under consideration: thus voluntaryconstruction, the housing of or gifts to theinstructor or teacher, etc. (p. 31).
While education would be clearly be-yond the reach of most parents were it notfor government subsidies, it remains asubstantial expense for the individualfamily. In addition to the costs alreadycited, parents must also dress and feedchildren; and in rural areas they foregothe advantage of having children assistwith farm work.
In Lagos, Nigeria, parents must payabout $20 annually for each child en-rolled in a municipal primary schooldespite the absence of a tuition fee. Ofthis total, $11412 is for books. The
14
balance pays for required school uni-forms, provides school lunches, andallows for contributions for variousschool activities (interviews with Lagosparents, checked with Nigerian FederalMinistry of Education). The Lagos semi-skilled working man, who earns perhaps$500 per year will encounter difficulty intrying to keep several children in school:To educate five children would require 20percent of his income. Despite the freeeducation scheme in Western Nigeria,primary education cost parents between$6 and $17 annually in Western Nigeria inthe mid -1960s (Callaway, 1967, p. 210).A more recent study has placed the costof a year of primary education at $24 to$45 and estimates that lower-incomegroups in Western Nigeria spend from 40percent to 60 percent of their incomes oneducation for their children (Briggs, 1971,pp. 49-50).
Costs seem to be even higher in ruralTogo. Midwives interviewed there said itcost them $80 a year to keep a child inprimary school. Of this $13 was forbooks, $50 for meals, $15 for uniforms,and $2 for school fees (Togo HealthMinistry Training Team, 1970). Evenallowing for the possibility that this figureincludes costs that the parent would haveto cover anyway, such as meals, primaryeducation is a major expense for thesewomen who earn approximately $200 peryear.
The increasing need for higher educa-tion has added to parents' educationalcosts. In most major cities of West Africanumbers of primary school graduates whocould fairly easily find jobs only a fewyears ago are unemployed. As school at-tendance becomes more widespread, moreeducation is needed to compete effectivelyin the job market. This means an addi-tional financial burden for parents. InGhana educational costs increase from$1.50 annually at primary school to $11at the secondary levelplus another $125for boarding fees unless the student ob-tains a government scholarship. In Ni-geria the least expensive secondary schoolsare those run by the federal government,and these cost $150 annually for tuition,boarding, and books.
In sum, the expansion of the educa-tional system in a country like Nigeriamay well be a primary cause for the grow-ing interest in limiting fertility. Educationhas been brought within the reach ofmillions of Nigerian parents prepared tomake a significant financial sacrifice fortheir children. Furthermore, educational
costs are likely to remain as influential inthe future as they have been in the past.Even if governments succeed in providingfree schooling, as the examples aboveindicate, school fees represent only aminor portion of parents' total educa-tional expenditures. They are certainlyminor compared to the additional expenseof supporting children through moreyears of schooling.
Family AllowancesAll 15 French-speaking countries of
West Africa have systems of family allow-ances for government employees andsalaried employees of private establish-ments. Each family covered receives cashpayments to cover medical costs duringpregnancy and at delivery. In addition,monthly payments based on the numberof children in a family are given to helpdefray the costs of raising children.
The particulars of family allowance pro-grams vary from country to country, butin most cases the amounts are substantialrelative to families' incomes. In general,the programs provide monthly paymentsduring pregnancy plus an additional pay-ment at delivery. The combined values ofthese vary from $25 (Chad, Mali) to $50(Senegal) and $70 (Guinea). Subsequentmonthly payments range from $1.25 achild per month (Chad) to $4 a child permonth (Guinea, Gabon). These paymentscontinue until the child reaches age 14 or15 years (12 in Guinea, 17 in Gabon),with provision for extension to 18 or 21years if the child remains in school, be-comes an apprentice, or is an invalid.
In Cameroon, Gabon, and Mali, thereis no restriction on the number of birthscovered. The more common pattern inother countries is to limit the deliverypayments to the first three births. Monthlysupport payments are given for an un-limited number of children in all coun-tries except Guinea (U.S. Department ofHealth, Education, and Welfare, 1967,passim).
The importance of these family allow-ances can be seen in Dahomey. InCotonou and Porto Novo, typists earnapproximately $45 per month, sales per-sonnel and telephone operators about $60per month (Dahomey Office of Statistics).Under the Dahomean system of familyallowances, such personnel would receiveprenatal benefits of $25 for each child plusan additional payment of $20 at the timeof delivery for the first baby, $10 for thesecond and third, and nothing for subse-quent children. Parents then receive a
9 219
monthly child support payment of $2.50per child (U.S. Department of HealthEducation, and Welfare, 1967, pp. 60-61).A Dahomean clerk would thus receiveabout $45almost a month's salaryinprenatal and delivery benefits for the firstchild and $35 for the second and third.If there were six children in the family, hewould receive $15 per month, increasinghis base salary by approximately one-third.
Some French-speaking West Africancountries also have a higher schedule ofbenefits for upper-level government em-ployees. Dahomey can again serve as anexample. Under this more lucrative pro-gram, professional and administrativegovernment officials (functionnaires) re-ceive $10 a child per month. The averagemonthly salary after taxes for these careermen starts at $150-200, and can rise to$250-300 after ten years of service (inter-views in Cotonou, Dahomey, February1970). A civil servant with six childrenwould draw about $60 per month infamily allowances, increasing his take-home pay by one-quarter. .
The value placed on this assistance bythose receiving it was demonstrated inDahomey in 1969. The government, facedwith one of its periodic financial crises.tried to cut back the $10 monthly "childsupport" payments to the $2.50 permonth allowed the average wage-earner.Government officials went on strike. Thegovernment eventually capitulated, retain-ing the $10 monthly payment level butlimiting the benefits to a maximum of sixchildren.
Since these particular family allowanceprograms apply only to government em-ployees and salaried workers in privateinstitutions, a relatively small part of thepopulation benefits from them. In Da-homey and Togo, for example, thoseeligible constitute only 10-15 percent ofthe labor force. In other countries of WestAfrica, the proportion is less than 5 per-cent, and in few countries does it exceed20 percent. Yet the eligibles are an im-portant group because they constitute themodern urban elite, or those who nor-mally would be concerned first with fer-tility limitation. Any program that slowsacceptance of family planning by thispace-setting group retards its acceptanceby the society as a whole.
No available data show whether or notfamily allowance programs actually influ-ence the recipients' reproductive behaviorand the society's population growth. Butinsofar as family allowances have any
IS
effect at all, it is presumably towardhigher rather than lower populationgrowth.
Governmental Policies that AffectRural-Urban MigrationGovernment development policies also
unintentionally affect migration. Almostall West African governments, for ex-ample, are promoting industrializationand modernization as fast as their re-sources will allow. There is a naturaltendency for industry to be located incities and towns. This concentration pro-pels people from rural to urban areas insearch of employment.
In addition, many government financialpolicies, by increasing the balanCe be-tween rural and urban incomes, have aneffect on migration. The common prac-tice of granting government marketingboards a monopoly on purchases of suchexport crops as cocoa, rubber, or ground-nuts (peanuts) at a price well below theworld market diverts the profils into thepublic treasury and lowers rural incomesrelative to urban incomes. One study ofNigerian Marketing Boards indicates thatfrom 20 percent to 30 percent of potentialproducer income never reaches the pro-ducer (Heilleiner, 1964, pp. 606-610).
In many instances, rural dwellers arecharged the same rate as urban dwellersfor services not provided in rural areas.Sir W. Arthur Lewis (1969) described thesituation in his 1968 Aggrey-Fraser-Guggisberg lecture at the University ofGhana:
The fundamental reason for this (currentdrift to the towns) is the big gap which hasnow opened up between urban wages and thefarmers' incomes. . . . An unskilled labourerin Lagos, in filll employment, is paid twiceas much as the average farmer earns. . . .Moreover, the gap is not confined to personalearnings, for there is an even bigger gap inpublic services. The governments levy taxeson the farmers and use the money to providewater supplies, medical services, secondaryschools, electricity, and transport not in therural areas but in the town. In these circum-stances any young man who stayed in thecountryside instead of migrating to the townsought to have his head examined (pp. 29-30).
Mother government policy that vastlyinfluences migration is educational de-velopment in rural areas. Professor Lewiscites the problems:
In Africa this swift transition [from limitedto widespread primary education in ruralareas] has raised the additional and morenegative problem of the effect of frustratedexpectations. For, in a country where only10 percent of children finish primary school,and there are virtually no secondary schools,any boy who finishes primary school is
assured of a clerical job in town at a highsalary. The primary school is thus establishedin the minds of parents and children as aroute to a white collar urban job. But when50 percent are finishing primary school, andin addition a new flow of secondary schoolboys are taking the clerical jobs and the jobsfor elementary school teachers, the primaryschool graduate can no longer find this kind ofwork. However, it takes time for people torecognize that times have changed. For someyears the primary school graduates continueto make a beeline for the towns, where theybecome unemployed. Depopulation of youthfrom the countryside, coupled with heavy un-employment in the towns, is now standard inmost African countries which have beenaccelerating primary education. One startsoff opening up schools in rural areas in thehope of having literate farmers, and findsinstead the effect is that the brighter mindssimply emigrate from the countryside intounemployment (pp. 27-28).
CONCLUSIONIn sum, few West African governments
influence population growth intention-ally, but all do unintentionally througheconomic and social development policies.Not enough is known about the quanti-tative effects of these policies to give aprecise evaluation of their overall impact.But the effect of health programs uponmortality has been so spectacular that theoverall impact is clearly to acceleratepopulation growth, even compensatingfor the possible offsetting effect of edu-cation.
Many more governments consciouslyseek to influence migration. These effortsappear to have succeeded in limitinginternational migration. But they havefailed to slow rural-urban migration, pri-marily because of the offsetting effect ofgovernment policies on such matters asindustrialization, education, and taxes.
For those who think rapid populationgrowth and rural-urban migration un-desirable, these unintended effects of de-velopment policies are unfortunate. How-ever, unfortunate population effects donot necessarily mean these policies shouldbe scrapped. To pay attention only to thepopulation effects of health policies, forexample, would be to ignore the in-estimable value of good health.
Improved public health may quitelegitimately be judged so beneficial thatit should be promoted despite an un-desirable population impact. But WestAfrican policy makers should be aware ofthe population implications of theiractions. For only when the impact ofdevelopment policies is recognized canneeded corrective measures be designedand implemented.
So far, West African policy makershave largely ignored the population as-pects of development. But their concernabout population is growing. Four yearsago, P.m would have predicted thatGhana and Nigeria would have adoptedpopulation policies or that Dahomey orThe Gambia would have commissionedstudies of their population situations. Thetrend is clearly with those who believethat population questions are too im-portant to be ignored.
REFERENCESAnnual Reports, 1967, 1968, 1969. Family
Planning Council of Nigeria.Briggs, Benoni. "Cost of education and cost
of living." Nigerian Opinion 7, no. 3(March 1971):48 -51.
Caldwell, J. 1968a. "Population policy: Asurvey of commonwealth Africa." In ThePopulation of Tropical Africa. Edited byJ. C. Caldwell and C. Okonjo. London:Longmans, 1968.
1968b. "Introduction." In The Popu-lation of Tropical Africa. Edited by J. C.Caldwell and C. Okonjo. London: Long-mans, 1968.
1968c. "The control of family size intropical Africa." Demography 5, no. 2(1968): 618.
Caldwell, John C. and Adenola A. lgun."The population outlook in Nigeria."Paper read at the Seminar on PopulationProblems and Policy in Nigeria, 22-27March 1971. Ife, Nigeria (Mimeo).
Callaway, Archibald. "Education expansionand rise in youth employment." In TheCity of Ibadan. Edited by P. C. Lloyd,A. L Mabogunje, B. Awe. London: Cam-bridge University Press, 1967.
Congo (Kinshasa), Haut Commissariat auPlan et a la Reconstruction Nationale.
Etude &Orientation pour la Relance Agricole.Prepared by the Institute de RecherchesEconomiques et Sociales, Universite Lou-vanium. Kinshasa: Bureau d'Etude duHaut Commissariat au Plan et a la Recon-struction Nationale, July 1966.
Dahomey, Plan de Developpement Econo-mique et Social du Dahomey, vol. I. Paris:Societe Generale d'Etudes et de Planifica-tion, 1961.
Daily Graphic, Accra. 23 January 1970: p. 1.de Buttet, C. "Bibliographic sommaire surles politiques demographiques en Afrique."Paris: Service de Documentation, InstitutNational d'Etudes Demographiques, 1970.
Dow, Thomas. "Sierra Leone." Country Pro-file. Population Council, September 196g.
Federal Republic of Nigeria. Second NationalDevelopment Plan 1970-74. Lagos: FederalMinistry of Information, 1970.
"Gabon." Afrique 1969. Paris: Jenne Afrique,1970.
220 16
Ghana. Population Planning for NationalProgress and Prosperity: Ghana PopulationPolicy. Accra: Government PublishingCorporation, March 1969.
Hallak, J. and R. Poignant. Les AspectsFinanciers de l'Enseignement dans les PaysAfricaines, d'Expression Francaise. Mono-graphies Africaines, no. 3. Paris: UNESCO,Institut International de Planifiaation del'Education, 1966.
Haute-Volta, Ministere du Plan. Plan-Cadrede Development, 1967-1970, volume 1.
Heilleiner, G. L. "The fiscal role of the mar-keting boards in Nigerian economic de-velopment, 1947-61." The Economic Jour-nal 74, no. 295 (September 1964): 606-610.
Lewis, W. Arthur. Some Aspects of EconomicDevelopment. The Aggrey-Fraser-Guggis-berg Memorial Lectures 1968. Accra:Ghana Publishing Corporation for theUniversity of Ghana, 1969.
Maboguaie, Akin L. "Regional mobilityand resource development in West Africa,"Centre for Developing Area Studies,
McGill University, Keith Canard Lec-tures, Series VI. Undated typescript.
Morning Post, Lagos. (anon.) 11 January1970, p. 12.
Niger. Plan Quadriemcd, 1965-1968, vol-ume I.
Nigeria, Ministry cf Economic Development,Guideposts for Second National Develop-ment Plan. Lagos: Ministry of Informa-tion, June 1966.
Nigeria, Federal Ministry of Education. Sta-tistics of Education in Nigeria, 1968. Seriesno. 2, volume 1. Lagos: Federal Ministryof Education Printing Division, 1966.
Pradervand, Pierre. "Obstacles to and possi-bilities of family planning in Francophonecountries of West Africa and the Republicof Congo." Paper presented to ExpertGroup Meeting on the Family Planningand Population Policies in Africa (Paris,6-8 April 1970). Development Centre ofthe Organisation for Economic Coopera-tion and Development (document no.CD/? /153), Mimeo.
Family Planning Acceptors in Lagos, Nigeriaby ROBERT W. MORGAN
Robert Morgan, Ph.D., is a senior research fellow in the Department of Com-munity Health of the University of Lagos College of Medicine. The author andthe Department of Community Health are grateful to the Family Planning Councilof Nigeria and its Chairman, Dr. Gabisu A. Williams, for making available theclinic record cards for analysis.
In Lagos, Nigeria, we made a compara-tive study of family planning clinic pa-tients and a sample of married womenaged 15-49. Our purpose was to look forindications of the kinds of persons whomight in time attend family planningclinics in Lagos and elsewhere in Nigeria,given further health education and theopening of additional clinics. Our majorfindings were as follows:
With respect to demographic indi-cators, marital status was found not to bea significant factor in clinic attendance.Age and parity were found to be highlysignificant, women aged 25 to 39 andwomen with more than three living chil-dren being represented in greater propor-tions among clinic attenders than in thepopulation at large.
With respect to socioeconomic indi-cators, virtually all clinic acceptors inLagos are from the lower socioeconomicclasses. Data on clinic patients, whencompared with data on a sample of theLagos population at large, suggest thatreligious and ethnic factors are not sig-nificant in family planning acceptance,
, r
and that the most important single factoris education. Those at all educationallevels were represented in proportionallygreater numbers among clinic attendersthan among the female population atlarge.
Among acceptors, over the entireperiod studied, 1958-1968, the IUD wasfound to be the preferred contraceptivemethod. However, in recent years, prefer-ence for the pill has emerged.
Lagos, because of its sizeover amillion people live in the metropolitanareathe level of modernization and in-dustrialization, and its cosmopolitannature, is probably unique in tropicalAfrica, and the findings here must beviewed with this in mind. Modern contra-ceptive practice is a new idea in Lagos andrelatively few persons have actually at-tended a family planning clinicapproxi-mately 20,000 women.
Sources of DataOur data are drawn from three surveys
conducted by the Department of Com-munity Health of the University of Lagos
Senegal, Deuxiime Plan Quadriennial de Di-veloppernent Economique et Sociale, 1965-1969.
Stillman, David and Jeane. "Policies affectingpopulation in Togo." Paper read at theSeminar on Population Growth and Eco-nomic Development, 14-22 December1969, Nairobi, Kenya, (Mimeo).
United Nations, Economic and Social Coun-cil, Economic Commission for Africa,"Demographic Content of African De-velopment Plans" (E/CN.14/POP/5, 30April 1969).
United States Department of Health, Educa-tion, and Welfare, Social Security Adminis-tration, Office of Research and Statistics.Social Security Programs Throughout theWorld, 1967. 1967.
West Africa. 1969. "Strangers in trouble."(anon.) West Africa no. 2739 (29 November1969).
West Africa. 1970. "When more jobs meanmore workless." West Africa, no. 2775(15 August 1970).
College of Medicine: a study of familyplanning acceptors, a survey of socio-demographic characteristics of the popu-lation, and a knowledge, attitudes, andpractice (KAP) survey of a subsample ofwomen. The acceptor survey coveredsocial and fertility characteristics of familyplanning clients over the period 1958-1968, the first decade of official familyplanning activity in the city. Data coverapproximately 7,750 patients registered asacceptors of some form of modern contra-ception at the clinics. A brief review of thehistory of family planning in Lagos willhelp to place these data in perspective.
Although the first family planning clinicin Lagos was opened in 1958, there was noformal family planning program until1964. The first clinic, in a midwife centeron Lagos Island, was open one evening aweek and offered diaphragms, condoms,and chemical contraceptivesthe pill andthe IUD were not available until theearly 1960s. During the seven-year period1958 -1964, about 600 contraceptive pa-tientsaround 50-100 per yearwereregistered. Reasons for these low attend-ance figures include lack of funds and ofgovernmental support, the inadequacy ofavailable contraceptive techniques, thewidespread faith of the Nigerian public intraditional contraceptive techniques, in-cluding the use of charms, abortifacients,
221 -
TABLE 1. Number of contraceptive wars registered by year, Lagos metropolitanarea clinics.
Clinic location
Lagos LagosYear island mainland Suburban Total1964 andearlier 91 9 1001965 398 154 5521966 771 356 1,1271967 1,146 853 43 2,0421968 1,388 1,938 97 3,423Totals 3,794 3,310 140 7,244° Excludes 121 palisnle for whom year of registration was not recorded and 350-400 patients registered
at Broad Street Clinic during 1955-1084, whose cards cannot be found.
and the traditional period of abstinencefrom sexual relations during lactation, andthe cultural value placed in many Africansocieties on large families.
In 1964, the National Council ofWomen's Societies, with headquarters inLagos, agreed to sponsor family planning.In November of that year, the advisorycommittee set up by the society was for-mally inaugurated as the Family PlanningCuuncil of Nigeria. Additional clinicsvTre opened in Lagos, and clinic registra-tions rose rapidly during the period 1965-1968, the totals for each year being 552,1,127, 2,042, and 3,423 respectively (seeTable 1).
This period coincided with the initia-tion of educational programs at variousinstitutions in the city, the beginning ofhouse-to-house field worker visits bygroups from the Department of Com-munity Health and the Family PlanningCouncil, and funding from foreign agen-cies. Though not included in the presentanalysis, registrations continued to riseduring 1969 and 1970 to 4,860 and 4,958respectively.
Currently, all clinics are operated orassisted by the Family Planning Council.Most meet one night weekly and arestaffed by one or two doctors and severalnurses recruited from various medicalfacilities in the city. The Council operates
one full-time day clinic on Lagos Islandand the University of Lagos operates twoprimarily research and training clinics onLagos mainland.
For the 1967 sociodemographic survey,30 sample blocks were chosen at random,each block containing a residential popu-lation of 400-500 persons. Selection wasbased on population densities in differ-ent parts of the city as indicated by the1963 census. The sample consisted ofevery permanent resident in each block atthe time of the survey, and a total of13,141 persons were thus enumerated, in-cluding 2,842 women in the 15-49 agegroup. Eighteen of the sample blockswere located in what was then the sepa-rately governed Federal Territory ofLagos, and 12 blocks were located inimmediately contiguous urban areas inwhat was then Western Nigeria. All partsof the metropolitan area now lie withinthe newly-created Lagos State. The 1968KAP survey covered marital and fertilitycharacteristics of a sub-sample of 729women aged 15-49 drawn from the 30blocks. Data on the total population ofpatients were drawn from actual clinicrecord cards and registers. The latter con-tained information on age, education, re-ligion, ethnic group, marital status andparity, in addition to clinical material in-cluding contraceptive method prescribed.
TABLE 2. Age distribution, Lagos metropolitan sample, clinic attendants,and 1963 federal census.
Age(years)
Lagos survey Clinics 1963 federal census
Number Percent Number Percent Number Percent15-19 533 18.8 117 1.6 24,996 17.020-24 599 21.1 769 10.6 37,491 25.425-29 577 20.3 2,027 27.9 32,242 21.930-34 456 16.0 2,265 31.1 21,530 14.635-39 307 10.8 1,328 18.3 14,003 9.540-44 227 8.0 841 8.8 10,630 7.245-49 143 5.0 129 1.8 6,560 4.4
Total. 2,842 100.0 7,276 100.0 147,452 100.0b
0 "Not Recorded" cases omitted In mast tables. Most of these omissions resulted from variations inclinic record cards in the earlier years. The population "bulge" in approximately the 20-34-Year age
grupgmay be sttributed to youthful in-mlgrants
surveysthe city who are seeking employment, and is re-
in all recent Lagos censuses and sample .
b Percent totals given as 100.0 throughout, although apparent totals may vary slightly due to rounding.
Demographic IndicatorsAGE
Table 2 presents the age distributionamong acceptors, among the Lagos sub-sample, and from the 1963 census for theformer Federal Territory. The census ma-terial is comparable only with the ma-terial for 18 of the 30 sample areas.Census data for Western Nigeria werepublished according to political divisionsas a whole and distinctions were notmade between urban fringe areas ofLagos and villages in the same divisionthat were not included in our sample.
Women aged 15-24 are underrepre-sented at the clinicsthat is, their propor-tion is lower than the proportion ofwomen in the same age groups in theurban sample. Women aged 25-39 areoverrepresented, that is, represented ingreater proportions than in the populationsample; women aged 40-44 are repre-sented in about the same proportion as inthe sample, and women aged 45-49 areagain underrepresented.
From these comparative data, one caninfer that women who become concernedabout family size do so at about age 25and this concern continues up to the early40s. Since Ohadike (1968) has shown thatthe mean age at marriage for Lagoswomen is about 20 yearsas comparedwith a lower age in the rural areasthemost surprising conclusion from thesedata is that these women apparently beginto recognize the need for child-spacingand family planning quite soon aftermarriage.
MARITAL STATUS
A review of the data (Tables 3-4) sug-gests that marital status is in general nota significant factor in clinic attendance.Data on clinic patients are compared withdata from the 1968 Department of Com-munity Health KAP Survey. During1958-1968, most Family Planning Councilclinics did not as a matter of policy renderservice to unmarried women. A numbnrof unmarried women may have obtainedservice at the clinics by reporting them-selves as married, and these women prob-ably reported themselves as marriedmonogamously. In reviewing Tables 3-4one should therefore begin by adding, inthe "clinics" and the KAP survey col-umns, the "never married" women tothose women married once or marriedmonogamously. Once this is done, onlyslight variations appear between clinicpatients and the sample population. Ofparticular interest in this connection is
222 18
Table 4, in which if this mathematicalcorrection is carried out the monogamyversus polyga...v factor disappears en-tirely, even to the irst decimal point.
Our finding that polygamously marriedwomen attend family planning clinics inthe same proportions (as readily) as domonogamously married women is cor-roborated by previous research: Ohadikein his 1964 survey in Lagos reports thatfertility differences between monogamous-ly and polygamously married women wereslight and probably not analytically sig-nificant (Ohadike, 1968).
The slight differences that appear, forexample with respect to "widowed"women in Table 3, can probably beattributed more to misreporting at theclinics than to actual social differences. Awoman can report any marital status shewishes when registering at a clinic, andshe may say simply that she is marriedrather than any other status. (The LagosKAP Survey is believed to be more accu-rate; it was an in-depth study carried outover a period of weeks, not only amongthe respondents themselves but amongother household members, neighbors, andfriends.) In summary, given the data inTables 3 and 4, relatively few differencesin clinic attendance patterns can bedemonstrated based on marital statusalone.
FERTILITY
Important differences between accep-tors and the sample population appear inthe data on fertility (Tables 5 and 6).Women with zero to two children arerepresented in lower proportions at theclinics than among the population;women with three children are repre-sented in almost the same proportion asin the sample of the population at large;and women at all higher parity levels arerepresented in the greater proportions inthe clinics. This implies that women inLagos who practice family planning beginto think about child-spacing after havinga third child.
This finding is particularly significantwhen one considers that the KAP survey,lace most other KAP surveys in WestAfrica, found an ideal family size of aboutsix children, including four sons and twodaughters, among the females inter-viewed (as did a parallel survey amongLagos males). If women begin thinkingabout resting or spacing their pregnanciesafter the third child, it would seem thatthey do not plan have children, andtheir actual prierred family size izev be
TABLE 3. Marital status, women aged 15-49, Lagos KAP survey and clinic attendants.
LagosKAP survey Clinics
Marital status Number Percent Number PercentNever married 56 8.3 12 0.2Married 583 86.9 6,313 95.8Divorced or separated 18 2.7 233 3.5Widowed 14 2.1 29 0.4Total 6710 100.0 6,587 100.0°Excludes 60-70 women who refused to give Information on marital status.
TABLE 4. Type of marriage, women aged 15-49, Lagos KAP surveyand clinic attendants.
LagosKAP survey Clinics
Type of marriage Number Percent Number PercentNever married 56 8.3 10 0.2Married monogamous 366 54.1 3,107 62.2Married polygamous 254 37.6 1,881 37.6Total 676a 100.0 4,998 100.0°Excludes 60-70 women who refused to give Information on marital status.
TABLE 5. Number of living children, women aged 15-49, Lagos KAP surveyand clinic attendants.
Number ofliving children
LagosKAP survey Clinics Expected
numberat clinics
RatioX100Number Percent Number Percent
0 123 18.2 83 0.9 1,330 51 135 20.0 297 4.1 1,462 202 140 20.8 708 9.7 1,521 463 108 18.0 1,208 18.5 1,170 1034 72 10.7 1,848 22.5 782 2115 50 7.4 1,532 21.0 541 2836 28 4.2 1,077 14.7 307 35178
124
1.80.8
455230
8.23.1
13244
3455239+ 2 0.3 97 1.3 22 441
Total 874 100.0 7,311 100.0
less than six. Women may cite the number"six" in attitude surveys either becausethey think this is the approved number orbecause they think this is the numbertheir husbands want. Certainly our im-pression at the Lagos clinics is that themajority of women, probably three-quarters of them, come in for contracep-tive service without their husbands'knowledge. On the other hand, Nigerianhusbands may also be becoming con-cerned about the expenses involved inraising a large family, and they too mayprefer fewer children than the stated ideal.Almost 70 percent of respondents, bothmale and female, in the KAP survey gaveeconomic factors such as the cost ofeducating a large family as one of "thebad things about having a lot of chil-dren," and this response pattern has beenreflected in most other West African sur-veys (Caldwell, 1968).
The data on number of live births(Table 6) parallel the data on number ofliving children (Table 5), but with some
223
interesting differences. One of the criticalproblems in West Africa, as in other partsof the developing world, is the high levelof infant and child mortality. It is knownthat many parents embark on repeatedpregnancies as a form of insurance, hop-ing that some children born will survivebut fearing that many will die. A centralissue among pediatricians concerned withfamily planning in West Africa thus hasbecome whether family planning shouldbe introduced on its own, so to speak, orwhether improved pediatric services anda demonstrated decline in child mortalitymust first be effected, after which motherswould then be prepared to practice preg-nancy spacing,.
While the numbers are small, Tables 5and 6 offer some support for the latterhypothesis. Among clinic acceptors, themedian figures are women with 4.8 livingchildren and 5.6 live births, indicating anaverage loss of one child per parent. Ifone examines expected numbers of clinicacceptors in each category, olges that
TABLE 6. Number of live births, women aged 15-49, Lagos KAP surreyand clinic attendants.
Number oflive births
LagosKAP survey Clinics Expected
numberat clinics
RatioX100Number Percent Number Percent
0 91 13.6 49 0.7 993 51 106 15.8 244 3.3 1,154 212 109 16.2 526 7.2 1,183 443 95 14.2 888 12.2 1,037 874 71 10.6 1,216 18.8 774 1575 72 10.7 1,283 17.3 782 1626 59 8.8 1,238 18.9 643 1937 33 4.9 846 11.6 358 2368 14 2.1 540 7.4 153 3539+ 21 3.1 494 8.8 226 219
Total 871 100.0 7,304 100.0
TABLE 7. Educational levels, woman aged 15-49, Lagos metropolitan sampleand clinic attendants.
Lagos survey ClinicsEducation Number Percent Number PercentNo formal 1,605 56.9 1,931 38.7Primary only 803 28.5 1,876 37.8Secondary and higher 411 14.6 1,180 23.7
Total 2,819 100.0 4,987 100.0
TABLE 8. Ratios of actual to expected numbers of clinic patients in eacheducational category.
Education Actual Expected Ratio X 100No formal 1,931 2,891 67Primary only 1,876 1,456 129Secondary and higher 1,180 640 184
for zero, one or two births the ratios ofactual to expected numbers are virtuallyidentical, whereas at higher levels womenwith more living children become pro-gressively better represented. This sug-gests that the death of a child does in factdeter a woman from family planning.
For the Lagos KAP sample, the mediannumbers are 2.6 living children and 3.3live births. The spread here is similar,that is, an average of 0.7 children lost permother as compared with a figure of 0.8for the clinic acceptors. One may suspectan underreporting of child deaths in bothsurveys, which are retrospective ratherthan prospective. At the same time, thefigures are not far out of line with othermore detailed studies conducted in theLagos Metropolitan area, and one mustalso recall that child mortality is Iilcely tobe lower in the city with its more sophisti-cated medical services than in the re-moter villages. Based on several urban andrural demographic surveys in Nigeria,this writer has estimated elsewhere an
Regrettably, direct data on children lost byeach mother are not available, because thecoding scheme used was developed by anotherresearcher to fit a specitic computer prom=and did not contain this information.
average of 5.7 surviving children and 1.6children lost per mother in Lagos and anaverage of 3.5 surviving children and 2.1children lost in rural Nigeria (Morgan,forthcoming). This would mean an aver-age of 7.3 live births per mother in Lagos,and, assuming that the average woman inthe KAP sample had passed halfwaythrough the childbearing period, figuresof 2.9 living children and 3.7 live birthswould be expected, not far different fromthe KAP figures, and suggesting only aslight underreporting of deaths.
Sodoecononsie IndicatorsEDUCATION
Most clinic patients are drawn from thelower classes in the Lagos urban az -.U.The relatively small wealthy elite in thecity, if they wish family planning, aremore likely to turn to private doctors forservice.
Table 7 shows that 37.6 percent ofclients have had some primary schoolingcompared to 28.5 percent in the generalsurvey. A total of 23.7 percent have ad-vanced as far as secondary school, com-pared to 14.6 in the general survey.
The influence of education on clinic
attendance becomes even clearer when, hiTable 8, the actual number of clinic pa-tients in each group is compared with theexpected number based on data from thesample population. Women with noschooling are represented in the clinicsonly 67 percent as well as in the popula-tion at large, whereas the correspondingratios for the other two groups are 129and 184 respectively.
With increasing emphasis on educationthroughout Nigeria, and with the highcosts associated with education beyondthe primary school level, this finding hasimportant implications for the future offamily planning in the country. Almost allfamily planning KAP surveys conductedin West Africa have demonstrated par-ents' concern about the cost of educatinga large family and the high value theyplace on large families.
RELIGIOUS AND ETHNIC DNTRIBUTION
Religious and ethnic factors are notsignificant indicators of family planningacceptance; although at first glance thedata would seem to indicate differences infertility behavior by religion. Accordingto Table 9, Muslims are underrepresentedat the clinics, Protestants are overrepre-sented, and Roman Catholics are repre-sented almost precisely in proportion totheir numbers in the population at large.
The "others" category comprises forthe most part the Separatist sects, that is,members of local churches such as Cheru-bim and Seraphim or Christ Apostolicwho have split off from parent Christianmissions. Normally these sects are groupedwith Christians, but the Family PlanningCouncil lists them in the "other" categoryin its record cards. We used the followingprocedure to estimate the proportion ofSeparatists among acceptors: in theLagos demographic survey, Separatistscomprised 7.2 percent of the total popula-tion, those who professed adherence totraditional African religions 0.4 percent,and members of "other" religions (pre-dominantly foreigners) 0.9 percent. Fromthis we inferred that Separatists com-prised about 85 percent of the "others"category in our clinic patient analysis.From this estimate, Separatists would ap-pear to be somewhat underreproented atthe clinics.
Based on the data in Table 9, amongChristians in Lagos, Protestants are bestrepresented, Roman Catholics are lesswell represented, and Separatists arereinvented least. One could interpretthese findings as follows. The proportion
224 20
of Roman Catholic clients is probablysmaller than that of Protestants becauseof the adverse influence of Papal au-thority: nurses in family planning clinicsin Lagos have reported that RomanCatholic patients have on several occa-sions come to have an IUD removed,giving as their reason the adverse state-ments of the Pope. Separatists are leastwell represented among Christians, andtwo reasons for this can be given. First,these splinter-group churches are presidedover by locally trained clergy known col-lectively as "Prophets," who in additionto their religious duties perform faithhealing similar to that in Christian Scienceand also have a considerable obstetricalpractice in their churches. The Prophetstend to be opposed on professionalgrounds to modern medical techniques,including family planning. Second, theSeparatist movement is more African thanforeign in its impetus, whereas moderncontraceptive techniques clearly are for-eign imports. The Prophets are in generalwell-informed and where we have en-countered them in our surveys, we havesucceeded in gaining their support for ourresearch and for family planning. This issociologically significant, since the con-version of a Prophet usually means theconversion of his many followers.
More difficult to explain, at first glance,is the apparent underrepresentation ofMuslims at the clinics. For most Muslimsin Lagos, religion does not hold the samepreeminent position in life that it does inthe north where Emirs and other religiousleaders have up to very recent times con-tinued to exercise much political andjudicial authority. Thus, it seemed un-liltely that their underrepresentation couldbe attributed to religious mores.
We hypothesized that it might be dueto ethnic origin and adherence to culturaltraditions that would preclude use ofmodern contraceptives. Virtually an LagosMuslims are members of the Yorubatribe, which made up some 86 percent ofthe sample pdpulation in the demographic
TABLE 9. Religion, women aged 15.49, Lagos metropolitan sampleand clinic attendants.
ReligionLagos survey Clinics
Number Percent Number PercentMuslim 1,497 53.3 2,144 42.4Roman Catholic 321 11.4 598 11.8Protestant 736 26.2 2,017 39.9Others 257 9.1 294 5.8
Total 2,811 100.0 5,051 100.0
° "Others" in most cases represent adherents to the so-ealled "Separatist" sects, see text.
TABLE 10. Religion, Yoruba women only, aged 15-49, Lagos metropolitan sampleand clinic attendants.
Religion
Lagos survey Clinics
Number Percent Number PercentMuslim 1,422 59.6 1,872 45.8Roman Catholic 148 6.1 298 7.3Protestant 630 26.4 1,828 39.9Other 188 7.9 288 7.0
Total 2,386 100.0 4,084 100.0
survey. Dr. Olatunde Oloko of the Uni-versity of Lagos has suggested that theYoruba Muslim group in Lagos might besociologically more closely knit thanother religious groups and hence slowerto accept any form of change.
To test the ethnic factor, the data forYoruba women alone were analyzed(Table 10). In the demographic sampleabout 60 percent of Yoruba women wereMuslim and the remainder mostly Chris-tian. For this single ethnic group the re-ligious factor remains approximately thesame as in the population at large, MuslimYoruba women being underrepresented atthe clinics, Roman Catholic Yorubawomen being represented in about thesame proportion as in the population atlarge, and Protestant Yoruba womenbeing overrepresented. Since most Sepa-ratists are Yorubas, their proportion inTable 10 is higher than in Table 9, but thedifference is artificial. It can thus be saidthat ethnic origin is not the explanatoryfactor.
When data on religion are cross-categorized by data on education, for boththe demographic sample and the clinic
sample (Table 11), the reasons for theMuslim underrepresentation become ap-parent:
(a) The majority of unschooled womenin both groups are Muslim (70.4percent of the Lagos sample and64.8 percent of clinic attendants).
(b) The majority of women who havehad any schooling at all, whetherprimary or higher, are Christians(69.6 percent of the Lagos sampleand 73.3 percent of clinic attend-ants).
(c) Within each religious grouping,women with schooling are repre-sented in greater proportions at theclinics than in the population attarp.
It would therefore appear that educa-tion is the predominant socioeconomicfactor in clinic attendance whether amongMuslims or among Christians and whetheramong one tribal group or another. Thefactors of religion and ethnic origin wouldthus seem not to be directly significant(with the exception as previously notedthat within the Christian subgroups,
TABLE 11. Edmiston, by religious alliiiiston, Lagos sample and clinic allandenb.
Women aged 15-49, Lagos sample Family planning clinic attendants
Education
Muslims Christians Total Muslims Christians
Number Percent Number Percent Nwnber Percent Number Percent Number Percent
Total
Number PercentNo formal
lyPrimary onSecondary and higher
Total
1,128 75.4 473 38.0 1,599 57.0 1,251 80.8 680 23.3 1,931280 18.8 520 39.8 800 28.5 804 29.3 1,271 45.8 1,87587 5.8 320 24.4 407 14.5 209 10.1 984 33.1 1,173
1,493 100.0 1,313 100.0 2,808 100.0 2,084 100.0 2,915 100.0 4,97
225
38.837.723.8
100.0
21
TABLE 12. Types of contraceptives prescribed, by yer, Lagos clinics.
Dia-Year Pills IUD phragm Foam Other Total1964 and earlier1965196619671968Not recorded
Total
1076
140142383
16767
(11.2%)
38287820
1,7232,561
835,492
(80.4%)
35153103
617220
444
(6.5%)
08
112047
1
87
(1.3%)
0756
210
39
(0.6%)
83531
1,0791,9523,084
1006,829°
(100.0%)
°Not including 538 patients registered who had service deferred to await evidence of non - pregnancy0.e. to await next menstrual periodi which Is the practice at most clinics. Experience has shownthat up to the time of recently tried free clinics, more than 90 percent of women who have paid theirregistration fee do return for service. Recent experimental free clinics have produced a greaternumber of acceptors and a higher proportion of defaulters. it should also be recalled that the pilland IUD were not Introduced until the fetter part of the 1958-1984 period.
TABLE 13. Ratio of IUD to pill patients, 1958-1970.
Year IUD PillRatio:IUD )( 1000
Pill1958-1964 38 10 3801965 287 76 3801966 820 140 5901967 1,723 142 1,2101988 2,561 383 6701969 3,331b 1,144b 2901970 2,880b 1,902b 150
°Other methods omitted.b Not including figures for two small suburban clinics In Ikejs and Shomolu.
Roman Catholics and Separatists are lesswell represented than Protestants.)
Contraceptives PrescribedTable 12 presents the distribution of
acceptors by method. During 1958-1968the IUD was used by the majority ofpatients (80.4 percent). In the years 1968-1970, however, the proportion of IUDusers fell off sharply, and by the end of1970 the pill was almost as prevalent inuse as the IUD (see Table 13). Rocco DePietro (1971) reviewing the years 1969-1970, has suggested that this trend repre-
sents a growing dissmisfaction with theIUD in Lagos. A broader review of thefull span of modern contraceptive use inLagos during 1958-1970 (Table 13) indi-cates another trend: in the earlier yearspills were used to a much greater extent;with the intensive health education cam-paigns beginning in 1966 which focussedon the IUD, the use of the IUD increasedto a high in 1967, the ratio of IUDs topills prescribed in that year being approxi-mately 12 to 1; and in succeeding yearsthe proportion of pill users has revertedto and more recently surpassed earlier
226
levels. One might surmise from these datathat the pill since its development has beenone of the preferred contraceptives inLagos.
ConclusionThree social and demographic factors
emerge from this study as predictors ofuse of family planning in Lagos: age,parity, and educational level. Women whoaccept family planning appear to be moti-vated to do so starting at age 25 and afterthree live births. These findings are sur-prising in view of KAP findings that theideal family size in Lagos is six children.The presence of any degree of educationwas found to increase the likelihood ofpractice of birth control.
REFERENCES
Caldwell, John C. "The control of family sizein tropical Africa." Demography 5, no. 2(1968): 598-619. (The Lagos KAP dataincluding ideal family size are reviewed.)
De Pietro, Rocco A., personal communica-tion 1971.
Morgan, Robert W. "Nigeria: Fertility levelsand fertility change." In Population Growthand Sodo-economic Change in West Africa.Edited by J. C. Caldwell. Forthcoming.
Ohadike, P. 0. "A demographic note onmarriage, family and family growth inLagos, Nigeria." In The Population ofTropical Africa. Edited by J. C. Caldwelland C. Okonjo. London: Longman (paper-back), 1968.
Oloko, Olatunde, personal communication1969.
Olusanya, P. O. "Nigeria: Cultural barriersto family planning among the Yorubas."Sturles in Family Flaming 1, no. 37 ("lulu-ary 1969): 13-16.
22
Use of Oral Contraceptives in Urban, Rural, and Slum Areasby MURARI MAJUMDAR, B. C. MULLICK, A. MOITRA, and K. T. MOSLEY
Mr. Majumdar is a statistician in the Indian Statistical Institute. Mr. Mullick isthe project director and Mr. Moitra the project statistician in the HumanityAssociation, an organisation engaged in welfare activities in West Bengal since1924. Mr. Mosley, formerly a consultant from the Ford Foundation in India, iscurrently at the Carolina Population Center. The project was carried out with thefinancial support of the Pathfinder Fund.
In Howrah District in India, betweenNovember 1968 and June 1970, we madea comparative study of the acceptabilityand use-effectiveness of oral contracep-tives in three clinics in different settings(urban, rural, and slum areas). Contra-ceptives were distributed by clinics free ofcharge. Field workers made repeated visitsto eligible couples to persuade them tcaccept at the clinics. The major findingsof the study were as follows:
In the urban area, where the programwas publicized before its initiation, amajority of acceptors attended the clinicprior to contact with field workers andmore than a quarter attended after onlyone contact. In the rural area, a largemajority came to the clinic after one con-tact with field workers, but in the slumarea, the majority of acceptors came onlyafter the second or third contact.
The cumulative continuation rates afterdifferent cycles of use computed by thelife-table method were highest in the ruralarea and lowest by far in the slum area.The slum acceptors showed higher drop-out rates through the entire period of use.
The cumulative continuation ratestended to be highest in the urban area forthe age groups 25-34 and for women withfour or five living children. The rates werehigher for literates but the difference be-tween literates and illiterates was substan-tial only in the slum area. The rates werealso higher in the slum area for womenwho joined the program after repeatedvisits by field workers than for those whocame after, at most, one visit.
For all segments combined, the pro-portionate increase in the cumulative con-tinuation rate was most marked in theslum area.
Although the ptIportion of acceptorscomplaining of side effects was quitelarge, particularly after the erg cycle ofuse, the number finally discontinuing be-cause of side effects was relatively small.The most important reason for discon-tinuation in the slum area was. related toproblems of use other than side effects,
whereas a majority discontinued in theurban area for reasons not connected withuse of the pill. In the rural area where thetotal number of women finally discon-tinuing was comparatively smaller, prob-lems of use other than side effects andother reasons were almost equally im-portant.
BackgroundIn November 1968, the Humanity As-
sociation started its Family Welfare Plan-ning Project for studying acceptors oforal contraceptives distributed by threeclinics in the district of Howrah, WestBengal. The first clinic, designated asurban, is in central Howrah town on theRiver Hooghly opposite Calcutta. Thesecond clinic, also in the town, is locatedin a slum area The third clinic functionsin a rural area about seven miles awayfrom Howrah and serves seven villages.
Each clinic served an area with a pop-ulation of roughly 10,000 according tothe 1961 census. Married women aged15-44 who were regularly menstruating
and eligible on medical grounds for takingsupplies of oral contraceptives were in-cluded in the project.
Two field workers, a man and a woman,visited each household in their respectiveareas to collect information about eligiblecouples and to persuade women to be-come acceptors of orals and visit clinicson specified dates. In case of a refusal orbroken appointment, the social workersmade two more attempts at persuasion.When a woman came to the clinic, sheunderwent a medical examination and inthe absence of contraindications, wassupplied with a cycle of orals.
The oral contraceptive first introducedwas Ovulen. Later a change was made toOvral, containing a lesser dosage of estro-gen, in conformity with the policy of theGovernment of India.
Characteristics of WomenCharacteristics of women contacted are
shown in Table 1. During the periodNovember 1968 to June 1970, 1,429 cou-ples were contacted in the urban area,1,488 in the slum and 1,304 in the rural.In the urban area all but two of thecouples were Hindus whereas in the slum,a majority, 54 percent, were Muslims and44 percent Hindus, and in the rural, al-most two-thirds were Hindus and theremainder almost all Muslims.
TABLE 1. Pomander distribution of eligible women by specified characteristics, forthree areas, November 19611-June 1970.
CharacteristicArea
Urban Slum RuralReligion
Hindu 99.9 43.7 64.8Muslim 0.1 54.0 35.0Other 2.2 0.2
Educational levelIlliterate 23.9 68.3 59.7Literate but below primary 32.4 10.5 19.9Primary 26.7 11.3 14.7Middle 16.4 9.9 5.9
Age (years)Under 20 7.1 9.9 13.820-24 20.3 19.3 19.825-34 50.0 45.1 41.135 and older 22.5 25.6 25.4
Number of living children0-1 30.7 33.8 28.82-3 32.4 34.7 27.44-5 22.9 21.9 27.16 or more 13.7 9.8 18.8
Total number of women 1,429 1,488 1,304
227 23
TABLE 2. Distribution of acceptors by number of visits by field workers, for threeareas, November 1968-June 1970.
Number ofvisits
Urban Slum Rural
N Percent N Percent N Percent
0 325 53.4 108 22.4 31 9.81 177 29.1 105 22.2 257 80.82 28 4.3 142 30.0 22 8.93 81 13.3 121 25.5 8 2.5
Total 609 100.1 474 100.1 318 100.0
The educational level of the eligiblewomen was comparatively higher in theurban area, but nearly a quarter of eventhe urban women were illiterate (thecriteria of literacy being ability to readand write a simple letter in any language)and another 33 percent, although literate,had not successfully completed the pri-mary four years of schooling. In both theslum and the rural areas, a majority ofthe eligible women was illiterate andmore than three-fourths had not com-pleted the primary stage of schooling.
The proportion of eligible women in theage range 20-34 was 70 percent in theurban area, 64 percent in the slum, and 61percent in the rural area.
The distribution by number of livingchildren shows that women have, on theaverage, the largest number of childrenin the rural area, fewer in the urban, thelowest number in the slum area.
Characteristics of AcceptorsTable 2 shows the distribution of ac-
ceptors by number of contacts with fieldworkers before acceptance. The figures inthe first row represent the women who
came to the clinic before any contact wasmade with them at home.
The total number of acceptors duringthe period November 1968 to June 1970was by far the highest-609-in the urbanarea; it was 474 in the slum and 318 inthe rural area. In the urban area, both thedistribution of oral contraceptives andfamily planning publicity programs hadbeen part of the Association's welfareactivities before the start of the project,although there had been no system ofhome visits or follow-up of acceptors. Thenew program received advance publicitythrough meetings and more than half ofthe total number of acceptors throughJune 1970 came to the urban clinic beforethe field workers had the opportunity tomeet them in their homes. A large ma-jority (80.8 percent) came to the ruralclinic after the first contact with fieldworkers, whereas among the slum dwellersa majority (55.5 percent) came only aftera second or third contact.
In the rural area, the proportions ofHindus and Muslims among acceptorswere practically the same as among eli-gible women, but in the slum area 70.7
TABLE 3. Percentage distribution of acceptors by specMed characteristics, forthree areas, November 1968-June 1970.
Characteristic Urban Slum Rural
ReligionHindu 100.0 70.7 64.8Muslim 24.9 35.2Other 4.4
Educational levelilliterate 41.0 46.4 70.4Literate but below primary 11.7 14.6 8.8Primary 30.8 18.6 15.7Middle and above 16.4 20.4 5.0
Age (year)Under 20 8.9 7.2 6.020-24 30.4 22.8 15.725-34 47.7 48.0 51.335 and older 13.0 22.0 /7.0
Number of living children0-1 17.7 14.0 8.92-3 39.2 33.3 29.64-5 28.2 28.8 32.46 or more 14.8 24.1 31.1
Total number of women 809 474 318
228
percent of acceptors were Hindus and 24.9percent Muslims although the latter con-stituted 54.0 percent of all eligible women.
In all three ar'as, most acceptors wereilliterate. In both the rural and urbanareas, not only were illiterate women pre-dominant among the accept .3 but theproportion illiterate was higher amongacceptors than among eligible women. Inthe slum area, however, the proportionilliterate was considerably lower amongacceptors than among the eligible popula-tion. The response in the slum areas maybe related more to the couples' prefer-ences in contraceptive methods and theavailability of methods than to socio-economic factors. We shall see later thatmany women in the slum area changedfrom the pill to condoms which had beenavailable in the area for a longer time.
The age composition and the distribu-tion by number of living children of pro-gram acceptors can be seen in Table 3.The acceptance rates (Table 4) per hun-dred eligible couples are higher in theurban area at ages under 25 than at olderages, the rate being maximum at ages20-24. Likewise in the slum area theacceptance rate is highest in the 20-24age group but the decline at older ages iscomparatively smaller. In the rural areathe rate is very low at ages under 25,rises steeply in the age group 25-34, andshows a slight fall at older ages. A markedvariation is also apparent between thethree areas in the acceptance rates bynumber of living children. In the urbanarea the rate is highest for women withtwo to five children and drops slightly forthose with six or more children. In theslum and rural areas, however, the ratesbecome progressively higher with increasein the number of living children and arehighest for women with six or morechildren.
Continuation Rates by AreaCumulative continuation rates were cal-
culated by the life-table method for allacceptors of pills during the period No-vember 1968 to June 1970. The data,except where otherwise stated, relate onlyto the first segment, that is, the experiencefrom the start of the use until the fastdiscontinuation or the cutoff date for con-tinuing users. A woman who failed toturn up on the due date or the day afterwas not given the supply for the cycle andwas treated as discontinuing if she visitedthe clinic. later or did not return evenwhen contacted by field workers. Thecases lost to follow-up were taken into
account for the intervals of use contrib-uted by them but were not treated asdiscontinuers. There was, however, nocase of a woman lost to follow-up in therural area. The women who were lost tofollow-up, were dropouts, or were not yetscheduled to return to the clinic at thecutoff date were credited with continua-tion up to the middle of the cycle follow-ing their last clinic visit.
The probability of continuation (Table5) was highest in the rural area, but themost outstanding difference is betweenthe slum area on the one hand and theurban and rural areas on the other. Thepercentage continuing for 12 cycles in theslum area is 30.6 compared to 61.1 in theurban and 66.6 in the rural area.
Differentials in ContinuationBY AGE GROUP
The cumulative continuation rates fordifferent age groups are presented inTable 6. The percentage continuing for aspecified number of cycles is highest forthe age group 25-29, and only a littlelower for 30-34. For all three areas con-tinuation rates at 12 cycles tend to behigher for women over 25 than for youngerwomen, although there is a dropoff afterage 35.
In the analysis of the cumulative con-tinuation rates by ages, it is apparent thatthe patterns in the first two cycles do notindicate clearly the pattern of differentialsthat emerges after longer intervals. Bycontrast, a major part of the differentialbetween the three areas at later cycles wasfound to have been due to differencesduring the first three cycles. Both the per-centages continuing the first three cyclesand those continuing for longer intervalsare found to be substantially lower in theslum than in the urban area, at all agegroups.
An interesting feature of the figures forthe rural area is the strikingly high cumu-lative continuation rate of women aged30-34. The percentage continuing the firstthree cycles comes to 95.5 in this agegroup whereas it varies between 81 and85 in other age groups. The divergencebecomes even wider after longer intervals:the percentage of rural women aged 30-34continuing 12 cycles is 87.5 compared to66.6 for all rural acceptors.
BY NUMBER OF LIVING CHILDREN
The cumulative continuation rates foracceptors by number of living children arepresented in Table 7. In all three areas,
TABLE 4. Number of acceptors per 100 eligible women by specified characteristics,for three areas, November 1988-June 1970.
Characteristic Urban Slum RuralReligion
Hindu 42.7 51.5 24.4Muslim 0 14.7 24.5LiteracyIlliterate 73.0 21.8 28.8Literate 33.2 53.9 17.9Age (years)Under 20 52.4 23.0 10.820-24 83.8 37.3 19.525-34 40.7 33.5 30.535 and older 24.5 27.0 25.9Number of living children0-1 24.8 13.2 6.32-3 51.7 30.2 21.34-5 52.6 41.1 29.16 or more 45.9 77.4 40.7
Total 42.7 31.8 24.4Number of women 1,429 1,488 1,304
TABLE 5: Cumulative continuation rate and dropout rate (percent per cycle) of firstchildren for three areas.
Cycle
Cumulative continuation rateInterval(cycles)
Dropout micaUrban Slum Rural Urban Slum Rural
3 79.0 55.9 85.5 0-3 7.0 14.7 4.86 70.5 42.3 76.4 3-6 3.6 8.1 3.59 85.3 34.8 69.1 6-12 2.2 4.6 2.112 81.1 30.8 66.6 12-15 2.0 2.4 2.615 57.5 28.4 61.3 15-18 1.6 0.618 54.8 ( ) 60.2 12-18 1.7 1.821 53.4 (60.2)Number of
women 609 409 318() Number of women at risk is under 10.( ) Number of women at risk Is under 20.a The dropout rats for a specified Interval was taken as the percentage decrease In the cumulativecontinuation rate over the Intern, divided by the number of cycles.
TABLE 6. Cumulative continuation rates (percent per cycle) of first acceptors oforals, by age, for three areas.
Cycle andarea
Age
<20 N 20-24 N 25-29 N 30-34 N 35+ N6 Cycle Rates
UrbanSlumRural
12 Cycle RatesUrbanSlumRural
71.9( )83.3
(52.1)
54
19
82.630.874.0
56.3( )
(62.9)
18510750
79.457.286.7
73.1( )54.55
17111794
73.541.191.8
67.4(31.1)87.5
12010869
64.342.974.4
50.1(6430.8)
.5
7910386
(*.) Number of women at risk is under 10.( ) Number of women at risk Is under 20.
TABLE 7. Cumulative continuation rates of first acceptors by number of livingchildren for three areas.
Cycle andarea Number of living children
Cycle 6 0,1 2 3 4 5 >5Urban 83.9 89.8 70.2 80.3 77.2 68.5Slum <25 (50.9) 41.1 57.1 (51.9) 44.9Rural (80.8) 73.2 80.1 73.3 79.1 75.3Cycle 12Urban 48.8 60.9 83.9 72.5 69.4 56.1SlumRural (88.1) (68.4)
( )(59.4) (78.0)
(35 .65.4
6)
() Number of women at risk is under 10.( ) Number of women at risk Is under 20.
229
TABLE 8. Cumulative continuation rate(percent per cycle) of firstacceptors of orals by literacyfor three areas.
Area andcycle
Cumulativecontinuation rate
Illiterate Literate
Urban3 75.3 81.46 67.3 72.8
12 58.3 63.118 51.9 56.7
Number of women 249 360
Slum3 48.0 62.86 32.1 51.3
12 (18.8) 40.8
Number of women 220 254
Rural3 83.6 90.16 73.7 83.0
12 85.6 69.518 59.3 62.9
Number of women 224 94
( ) Number of women at risk Is under 20.
the percentage continuing for a specifiednumber of cycles rises initially and thendeclines with progressive increase in thenumber of children.
The percentages, however, vary over aprogressively wider range with increase inthe number of cycles. Women with onechild or none show a particularly rapiddecline in continuation, with the percent-age continuing at least 12cyclesbeing46.6for these women as compared to 72.5 forwomen with four children. The percentagefor women with six or more children, onthe other hand, tends to decline moresharply than for those with five childrenafter later cycles.
Women with a small number of childrenmay adopt a contraceptive method forspacing their children or, if they adopt itfor family limitation, they may later de-cide they want additional children. Ineither case they are likely to use contra-ceptives for a limited period. Moreover,women with small families may not feelas compelled to practice family planningas those with large families and may not,therefore, be as conscientious in use ofcontraceptives. The continuation rate islikely to become progressively higheramong women with larger numbers ofchildren and, since older women havemore children, also with advancing age.It may be noted in this connection that 48percent of the women with six or morechildren were 35 years and older and 16
percent were 40 and older. At ages ap-proaching the end of the fertile periodwhen the family size is completed ornearly so, the expectation of additionalchildren even without contraception willgradually diminish and this is likely tooffset the tendency for the continuationrate to increase with larger numbers ofchildren and may explain the consistentlyhigh dropout rates for women aged 35and over and for those having six ormore children.
BY LITERACY
The cumulative continuation rates werecalculated for literates and illiterates (Ta-ble 8). The rate is substantially higher forliterates only in the slum area, althoughthe urban and rural rates for literates arealso slightly higher than those for illit-erates. The dropout rates appear to behigher in the slum area than in the urbanor rural for both literates and illiteratesbut it is the figures for the illiterates thatshow larger variations.
BY VISITS BEFORE ACCEPTANCE
A majority of the women who starteduse of the pill in the slum area came tothe clinic only after the field workerscontacted them for the second or thirdtime at home. Most urban and rural ac-ceptors came to the clinic after one con-tact or none. The continuation rates ofacceptors by number of contacts are,therefore, presented here only for theslum clinic. (The cumulative continuationrates are expressed as percent per cycle.)
CycleCumulative continuation rateNumber of visits
0-1 2-3
3 50.1 60.36 36.4 47.5
12 26.1 (36.0)
Number ofwomen 211 263
( ) Number of women at risk Is under 20.
The cumulative continuation rates ofwomen who came only after repeatedvisits by field workers and thus neededgreater persuasion were consistently high-er than the corresponding rates of thosewho accepted after one visit. At 12 cycles,the continuation rate for the formergroup was 36.0 compared to 26.0 for thelatter. Also, the dropout rates for womenwho came after at least two contacts werelower during early and later cycles.
TABLE 9. Cumulative continuation ratesof reacceptors of orals forthree areas.
Cycle Urban Slum Rural
3 54.6 44.9 62.96 43.8 36.1 39.49 35.8 (36.1) ( * )
12 (29.8) ( * )
Number of women 149 146 77
( ) Number of women at risk is under 10.( ) Number of women at risk Is under 20.
TABLE 10. Cumulative continuation rate(percent per cycle) for threeareas all segments com-bine&
Cycle
Cumulative continuation rate
Urban Slum Rural
3 92.8 73.6 97.96 88.8 63.0 95.79 84.5 56.1 90.4
12 77.6 47.4 87.715 72.1 38.5 86.218 66.7 (33.3) 85.121 83.6 (*) (85.1)
° Computation was based on Christopher Tletze."Intra-Uterine Contraception: RecommendedProcedures for Data Analysis." Studies in Fam-ily Planning. 18S: pp. 14, April 1987.
(*) Number of women at risk is under 10.( ) Number of women at risk is under 20.
RATES OF REACCEPTORS
The analysis thus far has been confinedto experience from the initiation of useof pills until the first discontinuation.Some of the women who discontinued thepill later resumed use. The field workersmade as many as three visits to the homesof discontinuers during consecutive cyclesto persuade them to come to the clinic onappointed dates. The continuation ratespresented in Table 9 relate to the experi-ence from resumption of use after thefirst interruption until the next discon-tinuation or the cutoff date for continuers.
The most noteworthy features of thesefigures are the low cumulative continua-tion rates in all three areas, and the muchsmaller differential than in the case of thefirst segment.
RATES FOR ALL SEGMENTS
Table 10 presents data for all segmentscombined, the final termination being thecutoff date for the study or the date ofdiscontinuance.
The percentages continuing for a spec-ified number of cycles are appreciablyhigher in all the three areas than those pre-sented earlier in Table 2 for t he first segment(see Table 11). The slum area continues tohave substantially lower percentages whiledifferences between the urban and ruralareas become progressively marked with
TABLE 11. Percentage Increase In cum-ulative continuation rates forall segments combined overthose for the first segment,for three areas.
Cycle
Percentage increase in thecumulative continuation rate
Urban Slum Rural
3 17.5 31.7 14.56 25.9 48.9 25.39 29.0 61.2 30.8
12 27.0 54.9 31.715 25.4 35.6 40.618 21.7 41.421 19.1
longer intervals. The percentage continu-ing 18 cycles comes to 66.7 in the urbanarea compared to 85.1 in the rural, thecorresponding figures for the first segmentbeing 54.8 and 60.2 respectively in theurban and rural areas.
TABLE 12. Number of oral acceptors and percent complaining of side effectsfor three areas.
Urban Slum RuralNumber of Percent Number of Percent Number of PercentCycle acceptors complaining acceptors complaining acceptors complaining
1 609 18.6 469 27.1 318 29.22-6 437 3.4 212 14.3 239 13.27-12 262 2.5 72 12.7 134 13.013-18 134 2.2 20 14.9 67 8.7
COMPLAINTS AND DROPOUT
Very few women dropped out becauseof side effects, but many reported sideeffects. Almost one-fourth of the acceptorscomplained of adverse reactions after thefirst cycle of use of orals (Table 12). Theproportion was comparatively smaller inthe urban area, 18.6 percent, the slum andrural percentages being 27.1 and 29.2 re-spectively. After continued use, the pro-portion of experiencing side effects rela-tive to the number of users dropped to alow level in the urban area, but in theslum and rural areas, although there wasa marked reduction during cycles. 2-6,there was no significant trend thereafter.Percents of those experiencing side effectsby specified type of side effect are shownin Table 13.
The intensity of discomfort that maylead one to report side effects varies amongindividuals and the large difference be-tween proportions complaining in theurban clinic on the one hand and the slumand rural clinics on the other may beattributed, at least partly, to differencesin personal reactions in different environ-ments. Actually, the side effects could not,in most cases, be severe as only 11 womenin the urban, 17 in the slum, and 2 in therural area discontinued the pills becauseof side effects (Table 14). Also nausea andvomiting, a major complaint after the firstcycle of use in all the three clinics,, andbreast discomfort in the slum area werereported by only a few after continueduse. By comparison, headaches, dizziness,and break-evough bleeding continued tobe reported, although by relatively fewerwomen, in later cycles. The few womenwho reported leg pain were not found
TABLE 13. Oral acceptors by type of side effect, lot three areas.
Percent with specified type of side effecta
Areaandcycle
Numberofacceptorspercycle
Head-acheanddizzi-ness
Break- Nausea Breast Abdom-through and discom- inalbleeding vomiting fort pain
Legpain
Allergicreaction Anxiety
Urbanb1 609 46.0 33.0 58.0 6.0 4.02-6 . 437 4.0 8.4 2.2 2.8 1.6 0.67-12 262 1.2 5.8 0.3 1.7 0.3 0.213-18 134 1.5 2.3 0.3 0.5 0.5Slumb1 469 50.0 35.0 41.0 10.0 28.0 9.0 1.0 6.02-6 212 12.2 8.0 2.0 1.4 9.4 3.0 0.67-12 72 3.7 2.8 0.5 0.2 2.5 0.7 0.513-18 20 0.8 1.2 0.2 0.2 0.8
Rural!)1 318 40.0 37.0 31.0 1.0 1.0 2.0 3.02-6 239 10.4 20.2 4.0 0.4 1.2 1.2.7-12 134 5.3 12.7 0.8 0.3 0.8 0.2 0.513-18 67 3.3 4.0 0.3 0.2 0.3
° Percents sum to more than 100 because some respondents reported more than one side effect.b There was also one case of pigmeMation in the sixth cycle in the urban area and one such case inthe third as well as in the fifth cycle In the rural area.
TABLE 14. Number discontinuing the oral by reason for discontinuation,for three areas.
Reason Urban Slum RuralSide effects 11 17 2Other problems of use 9 135 11Husband's objection 4 7 6Physician's advice 4 22 1Adverse publicity 2Religious reasons 2Lack of confidence in drug 1
Switched to another method 106Other reasons 35 9 13Not needing contraception 8° 3b
Desire for a child 10 4 6Unplanned pregnancy:Using pills irregularly 8 2 3After discontinuation 1 3Others 8 1
Reasons unknown 8 4Total 63 165 26Lost to follow-up 30 21° Vasectomy -4; tubal lipation-2; separation or death of husband-1; husband sick or away-1.b Vasectomy-1; separation or death of husband-2.
after examination by physicians to havedistinctive symptoms of thromboembolicdisease.
The reasons for discontinuation pre-sented in Table 14 were ascertained onlywhen the women did not resume use ofthe pills even after repeated visits by fieldworkers. Out of 11 women in the urban
area and 17 in the slum who dropped outbecause of side effects, eight did so in eachcase during the first two cycles.
In the urban area, nine women discon-tinued because of other problems of use,mostly husband's objection or "physi-cian's advice." It was found that in allthe three areas, when the women cited
231 27
physician's advice, the physicians con-cerned were not qualified medical prac-titioners.
In the slum area out of a total of 165women who discontinued the pills, 106had switched to the condom for reasonsnot clearly stated. It was noted earlier thatthe continuation rate in the slum areawas much lower than in the urban andrural. It appears that in a large majorityof cases of final discontinuation in theslum area, the women had been undersome influence to substitute the condomfor the pill. Condoms are distributed freeby local medical practitioners and havebeen available since mid-1969. The phy-sician's advice had also a greater impacton discontinuation in the slum area thanin the urban and rural: 22 women dis-continued the pills for this reason in the
slum area as compared to four in the urbanand one in the rural area.
The other specified causes, not relatedto problems of use, account for a muchlarger number of cases of discontinuationin the urban area than in the slum orrural. Generally they fell into three cate-gories: first, women who did not need con-traception because they or their husbandshad been sterilized or because their hus-bands were not present; second, womenwho desired additional children; andthird, those who had unplanned pregnan-cies during irregular use of pills. (Oneurban and three rural women reportedunplanned pregnancies during follow-uphome visits after they had already discon-tinued the pills.) In addition eight womenin the urban and one in the rural area whoreported they were pregnant only after
using the pills for some time were sus-pected to have concealed the fact of theirpregnancy prior to acceptance of the pills,having imagined that these could be usedas abortifacients.
Conclusion
The question that is of primary impor-tance in this study is whether Indianwomen, particularly those who are in therural areas and are illiterate, will acceptthe pill and whether they will continueusing the pill over a period of time, Thedata show that oral pills can be an accept-able method of contraception for Indianwomen. The high continuation rate in theurban and rural areas is also an importantfinding in its favor.
THE POPULATION COUNCIL245 Park Avenue. New York. New York 10017
The Population Council is an organization established in 1952 for scientific training and study in the fieldof population. It endeavors to advance knowledge in the broad field of population by fostering research, train-ing. and technical consultation and assistance in the social and biomedical sciences.
BOARD OF TRUSTEES
John D. Rockefeller 3rd. ChairmanNew York City
Bernard BerelsonPresident, The Population Council
Sissela BokCambridge, Massachusetts
Detlev W. BronkPresident Emeritus, The Rockefeller University
John C. BullittShearman & Sterling
Mary I. BuntingAssistant to the President, Princeton University
Robert H. EbertDean, Harvard Medical School
Roswell L. GilpatricCrawl,. Swains & Moore
Caryl P. HaskinsWashington, D. C.
A. Leon Higginbotham, Jr.Judge, U. S. District Court for theEastern District of Pennsylvania
W. David HopperPresident, International Development ResearchCentre, Ottawa
Cordelia S. MayPittsburgh, Pennsylvania
Donald H. McLean. Jr.President, Lahey Clink Foundation, Inc.
John T. Noonan. Jr.Professor of Law,University of California, Berkeley
Frank W. NotesteinPresident Emeritus, The Population Council
Theodore W. SchultzProfessor of Economics. University of Chicago
U ThantFormer Secretary-General, United Nations
Gilbert F. WhiteDirector, University of ColoradoInstitute of Behavioral Science
Salvador Zubiran, M.D.Director, Institute. Notional de laNutrition, Mexico
232 28
