Ceylon: continuing practice of contraception by acceptors of oral contraceptives and intrauterine devices in a field programme*

Bull. Org. mond. Sat 1973, 48, 639-647 Bull Wid Hlthi Org. Ceylon: continuing practice of contraception by acceptors of oral contraceptives and intrauterine devices in a field programme* NICHOLAS H. WRIGHT 1 & TERRENCE PERERA2 A follow-up study of oral-contraceptive and intrauterine-device acceptors in the early years of the Ceylon family planning programme was undertaken to determine the effect ofa 1968 policy decision to cut the per cycle price of oral contraceptives by half and to allow nonclinical prescription and distribution of the pills by trainedfield midwives under medical supervision. While pill acceptance increased, study findings suggest that continuation did not. On balance, the intrauterine device proved much more effective than oral contraceptives in preventing unplanned pregnancies, reflecting the higher continuation rates of IUD acceptors. Although the availability of both methods clearly contributed to expansion of theprogramme, the demographic and administrative implications ofinitial choice ofmethod, and the finding that pregnancy rates for pill users are not lower than for IUD wearers, suggest the needfor more realistic counselling by clinic andfield staffat the time of choice. When family planning officially became part of the Maternal and Child Health Programme of the Ceylon Ministry of Health in 1966, among the first of the fifteen health areas to receive contraceptive services were Galle and Kalutara, two provinces on the south-west coast. A pilot clinic established in 1958 at Bandaragama in Kalutara under the Sweden-Cey- lon Family Planning Project had supplied intrauterine devices and pills locally by 1964. From early 1966 these methods, as well as other traditional ones, were available in selected health centres, clinics, and hospi- tals in both areas. Although the programme first emphasized the IUD, acceptance of oral contraceptives increased, particularly in the last half of 1968. In April 1968 the price per cycle was reduced by half to US$0.13 and policy was changed to allow the Ministry's * From the Family Planning Bureau (now Family Health Bureau), Ministry of Health, Government of Ceylon. Subse- quent to the completion of this work the name of Ceylon was changed to Sri Lanka. 1 Formerly Resident Adviser, Population Council-Ford Foundation Project. Now Representative, The Population Council, and Medical Adviser, National Family Planning Program, Ministry of Public Health, Thailand. ' Formerly Director, and Assistant Director of Health (Maternal and Child Health). Now Regional Adviser (MCH), WHO Regional Office for South-East Asia, New Delhi, India. trained field midwives,3 under supervision of their medical officer, to prescribe oral contraceptives on the basis of medical history and to carry continuing supplies to users' homes. Earlier, programme oral- contraceptive supplies could be prescribed only after physical examination by a medical officer and were available only at selected clinics. By 1969, almost 50% of new acceptors in the programme chose oral contraceptives and about 35% the IUD, a reversal of the pre-1968 acceptance pattern. Almost all the increase in new acceptors, from 48 164 in 1968 to 56 787 in 1969, was due to new oral-contraceptive acceptors. Follow-up studies ofIUD acceptors in three health areas in 1968 and early 1969 had shown first method, all-segment continuation rates at 24 months of 80, 79, and 70 per 100 first acceptors, respectively (Wright, unpublished data,4 1973). Preliminary impressions and a small follow-up study using clinic records alone suggested that oral-contraceptive continuation rates were lower, giving cause for con- " Field midwives in Ceylon have 8-10 years of basic schooling, one year of professional training with 6 months, field practice under supervision, and a 1-week family planning training course. ' To be published in the Population Council Country Report Series. 3061 639 NICHOLAS H. WRIGHT & TERRENCE PERERA cern. It was not clear that the increase in new accep- tors and the proportion accepting oral contraceptives represented a corresponding increase in continuing effective use of contraception. Two studies were planned to attempt to clarify the situation. The first, a follow-up study of 1966-67 oral-contraceptive and IUD acceptors at the same Galle and Kalutara clinics, is described in this paper. The second, a follow-up study of only oral-contraceptive acceptors after mid-1968, was designed to see whether initial prescription and field distribution by midwives, as well as the reduction in price per cycle, improved continuation among the increasing numbers of first acceptors. Unfortunately, the second study could not be carried out as planned, but efforts to define continuation among more recent cohorts of oral- contraceptive acceptors in an area culturally similar to Galle and Kalutara are referred to. SAMPLE AND METHOD The six clinics in the Galle-Kalutara area offering both IUDs and oral contraceptives during 1966 and 1967 were identified and a systematic sample of acceptors drawn from clinic records. A sampling fraction of t in 20 for IUDs and 1 in 2 for oral contra- ceptives yielded 658 cases, about equally divided between the two methods. There were 45 cases excluded from the sample because of duplicate records (17), pregnancy at the time of acceptance (13), failure ever to insert an IUD or to take oral contraceptives from the clinic (8), first acceptance in years other than 1966 and 1967 (4), acceptance for therapeutic reasons (2), and other reasons (1). The remaining sample of 613 cases con- sisted of 291 IUD and 322 oral-contraceptive accep- tors. Follow-up efforts began in mid-1969. In the first phase, field midwives were asked to contact sampled cases living in their areas, requesting them to return to the clinic of acceptance on a fixed date for interview. A compensation was offered for the special travel by the former acceptors and field midwives. Slightly less than 60% of the sample were interviewed following the first request and another 10% during a second attempt. Home interviews based on the clinic record and the comments of the field midwives who had just sought to contact these women subsequently located an additional 20% of the sample. At the conclusion of the survey, 539 of 613, or 88% of the sample, had been interviewed. Of the remaining 74 acceptors in the sample, 49 had moved temporarily or permanently from the Galle- Kalutara area without leaving a new address, 22 were unknown (incorrect name at acceptance or unknown to field midwives newly assigned in the area), 2 re- fused to be interviewed, and 1 had died. In many instances, continuing information available from these 74 women on the clinic record card was used to complete the questionnaire up to the last contact between the client and clinic or field midwife. The basic questionnaire used for the study was the Population Council's " Follow-up record for studies of contraceptive use-effectiveness" (Tietze, 1968). A supplement to the questionnaire was added to record information on fertility and contraceptive practice before acceptance, education, desire to limit or space children, and additional data on oral contraceptive use. RESULTS Acceptor characteristics Oral-contraceptive acceptors in the sample tended to be younger and have fewer living children than IUD acceptors (Table 1). The differences, however, were small. The mean age at acceptance for oral- contraceptive and IUD acceptors was 29.6 and 30.0, and the mean number of living children 4.1 and 4.5, respectively. Among the sampled women who were interviewed, prior fertility was high and did not vary significantly by method accepted. In the 3 full calendar years preceding the year of acceptance fertility rates were 400 and 416 per 1 000 women for oral-contraceptive and IUD acceptors, respectively. In 1963, marital fertility among all Ceylonese women 15-49 years of age was estimated to be 239 per 1 000 (Wright, 1968). However, if married Ceylonese women in 1963 had had the same age distribution as the younger acceptor sample, their fertility rate would have been 289 per 1 000, thus narrowing the difference. In so far as many of the women had delivered recently, their pre-acceptance fertility was exaggerated. Had all the interviewed acceptors been married and having children for a full 3-year period before the year of acceptance, however, their prior fertility might well have been higher. Although approximately two-thirds of the inter- viewed sample had at the time of acceptance an open interval of less than a year, i.e., were within a year of their most recent delivery or abortion, many acceptors of both methods, de pite previously high fertility rates, delayed accepting and risked another pregnancy. As Table 2 shows, IUD acceptors 640 CEYLON: CONTRACEPTION IN A FIELD PROGRAMME Table 1. Numerical and percentage distribution of acceptor sample by age and number of living children Method accepted Characteristic Oral IUD no. % no. % age under 20 5 2 3 1 20-24 67 21 43 1 5 25-29 100 31 101 35 30-34 77 24 80 27 35-39 51 16 47 16 40+ 18 6 17 6 unknown 4 1 - - no. of living children 0 3 1 - - 1 29 9 12 4 2 51 16 41 14 3 67 21 54 19 4 48 15 58 20 5 34 11 40 14 6+ 85 26 86 30 unknown 5 2 - - total 322 100 291 100 had shorter open intervals even though postpartum lUD insertion was not available in 1966-67. No substantial differences were observed among the interviewed cases by method, however, in the pre- ceding closed interval, i.e., the time interval between the most recent delivery or abortion and the one Table 2. Cumulative percentage distribution of inter- viewed sample by open interval and method accepted Open interval Oral IUD (months) 3 18 31 6 33 51 12 60 69 18 75 80 24 83 88 immediately before it. For the 513 of the 539 interviewed acceptors with more than one pregnancy, the mean closed interval was 27 months. Less than 10% and slightly more than 50% of these women had closed intervals of less than 12 and 24 months, respectively. Among the interviewed group, 14% had never attended school. A majority, 56%, had completed the fifth standard or a higher level. Younger women and oral-contraceptive acceptors had completed more standards than older women and IUD acceptors. Twenty-six percent had practised family planning before acceptance, mostly by means of periodic abstinence, a term that the interviewers believed was often taken to mean absence. The number practising withdrawal was probably understated. Only one- fifth of past users, i.e., 5% of the sample, had used oral contraceptives or an IUD. It seems clear that among the 1966-67 acceptor sample there was mini- mal substitution and considerable initial technologi- cal improvement in family planning practice. Past use was higher among older women and, except for women of the highest parity, related to the number of living children at acceptance (Table 3). There was no relationship between pre-acceptance practice status and education. A large proportion (60%) of interviewed women wanted no more children. The proportion increased steadily with age and number of living children, reaching 43% for those with 3 and 70% for those with 4 living children. A higher proportion of loop acceptors (66 %) than of oral-contraceptive acceptors Table 3. Percentage of interviewed women practising family planning at any time before acceptance, by age and number of living children Characteristic % Age under 25 22 25-29 16 30-34 33 35 and over 36 no. of living children 1-2 15 3-4 31 5 and over 27 641 NICHOLAS H. WRIGHT & TERRENCE PERERA Table 4. Net cumulative continuation rates per 100 acceptors by first method, and by all methods accord- ing to first method, up to 24 months Months Method accepted 6 12 18 24 first method oral 49 40 35 30 IUD 86 80 75 68 all methods by first method oral 65 58 52 47 IUD 91 84 80 75 (55 %) wished to limit their number of children, but these differences reflect in part the age and parity differences shown in Table 1. Continuation rates andpregnancy rates First method IUD continuation rates were double those for oral contraceptives at 12 months, and more than double at 24 months (Table 4). Although the gap is somewhat reduced when all-method continuation rates are considered, i.e., allowing for the adoption of a second or later method when the first is disconti- nued, a considerable difference remains at 24 months. Acceptors aged 30 years and over continued using either first method longer than younger acceptors. This was also true for all method continuation rates, as seen in Table 5. Women with more living children at acceptance also continued use of the first or another method longer than those with fewer chil- dren. While oral-contraceptive acceptors with higher completed educational levels continued longer than those with lower ones, the reverse was true for the IUD acceptors in both continuation categories. How- ever, none of the comparisons within these methods were significant. The main reasons for termination of the first method were defined as pregnancy, change ofmethod, planning pregnancy, and other discontinuation. Preg- nancy rates during use were similarly low among oral-contraceptive and lUD acceptors. Oral-contra- ceptive and older acceptors changed methods more than lUD and younger acceptors, respectively. Younger women, especially oral-contraceptive accep- tors, were much more likely to be planning another pregnancy. " Other " discontinuation was much Table 5. Net cumulative continuation rates per 100 acceptors at 12 months by age, number of living children, and education All methodsFirst method by first method Characteristic by_first_method Oral IUD Oral IUD age 15-29 38 75 53 78 30-49 42 85 63 90 living children 0-3 37 74 55 79 4 or more 42 84 60 87 education below 6th standard 34 83 51 87 6th standard or higher 40a 75 61 80 a Based on less than 50 woman months of experience. higher among acceptors of oral contraceptives, particularly those aged under 30 years, the main reason being side effects (60 %). Clinic-related factors such as distance of the clinic from the client's home, unavailability of supplies at the clinic (breakdown in pill supply line), and lack of money to purchase pills accounted for 23 %. The balance discontinued often for vague general health considerations, but sometimes because of specific objections on the part of the husband. " Other " discontinuation among IUD acceptors was chiefly due to medical removal and expulsion without reinsertion. Very low medical removal rates and high reinsertion rates among women experiencing expulsion mainly accounted for the high IUD continuation rates. Characteristics of the acceptors other than age, living children, and completed educational level were associated with continuation of the first and of all methods. Table 6 shows that the acceptor cohort of 1967 did not continue contraceptive practice as long as did 1966 acceptors. Acceptors with shorter open intervals tended to continue longer, as did those who intended to limit family size. All rates for intention to space or limit births were significantly different at a level of P <0.05. Women who had practised family planning before acceptance had higher all- method continuation rates than those women never practising before, but the differences were not signi- ficant. Finally, continuation rates were significantly 642A CEYLON: CONTRACEPTION IN A FIELD PROGRAMME6 Table 6. Net cumulative continuation rates per 100 acceptors at 12 and 24 months for combined first methods and for all methods by selected characteristics First method All methods Characteristic months months 12 24 12 24 year accepted 1966 59 49 73 64 1967 59 46 68 57 open interval (months) 1-6 64 53 73 63 7 or more 54 43 67 57 previous practice of family planning yes 55 47 71 63 no 59 47 69 58 intention a spacing of births 50 36 61 46 limiting of births 63 54 75 68 place of interview a health centre, clinic, hospital 62 51 73 63 home, other 43 34 b 57 48 a Rates between categories significantly different (P <0.05) at all time periods. b Based on less than 50 woman months of experience. higher (P <0.05) for those acceptors interviewed at the health centre, clinic, or hospital as opposed to the women's home or some other place. There were profound differences in post-acceptance pregnancy rates among acceptors by first method and by all methods by first method. The gross cumulative pregnancy rates reported in Table 7 at 12 and 24 months reflect all pregnancies during use of the first method (use-effectiveness) as well as unplanned pregnancies with other methods and after termina- tion of all contraception (extended use-effectiveness). For oral-contraceptive acceptors, the pregnancy rate per 100 users at 12 months is 3. When unplanned pregnancies during the use of another method and after terminating all contraception are included, the rate is 24. The trend for IUD acceptors is similar, although the gap is much smaller and both rates are lower than the comparable rate for acceptors of oral contraceptives. At 24 months from acceptance, 44% Table 7. Gross cumulative pregnancy rates per 100 users, at 12 and 24 months, during use of first method only and by all methods a according to first method Months Method in use 12 24 first method oral 3 4 IUD 1 3 all methods a oral 24 44 IUD 8 17 a Including unplanned pregnancies with other methods and after termination of all contraception. of the latter had been or were pregnant. For IUUD acceptors the figure is 17%. These two percentages summarize the relative demographic effectiveness of the two methods in this study. Early discontinuation and interruption Answers to the supplementary questionnaire re- vealed that 7% of the acceptors of oral contracep- tives who were interviewed never actually used the pills after undergoing examination and purchasing them at the clinic. Of the 18 women in this group, 7 heard unfavourable rumours and 6 met with objections from their husbands. The other 5 gave nonspecific reasons. Answers to other questions in the supplementary schedule suggested that those acceptors who had not begun taking the pills as instructed after leaving the clinic still had not taken them up to the date of interview. Only 2 of the 274 interviewed pill acceptors delayed a month or more before starting. On the other hand, 58 acceptors, or just over 20%, discontinued after the first cycle. It was asked if pill taking had been interrupted at any time and, if so, why. Women who never started taking their pills (18) and women taking only one cycle (58) were excluded. Of the remaining 198 inter- viewed women, 30 reported interruptions. The main reason given was to allow minor side effects to resolve. Only a small number cited clinic-related reasons, such as absence of supplies at the clinic or distance. These findings are illustrated in the accom- panying diagram. -1 A NICHOLAS H. WRIGHT & TERRENCE PERERA interviewed acceptors of oral contraceptives acceptors (N = 274) 11 1 never began use began use (N = 18) after delay (N = 2) began use at acceptance (N = 254) took more than 1 cycle (N = 198; I (took only 1 cycle N = 58) Table 9. Net cumulative continuation rates per 100 oral contraceptive acceptors at 6 and 12 months for first and all methods by usual source of supply " First method All methods Usual source months months 6 12 6 12 clinic and/or private pharmacy midwife at client's or midwife's home b 56c 41 c 69C 57 69 60 87 84 clinic and/or midwife t 87 c 74 c 90 C 80 interruptions while taking (N = 30) continuous use while taking (N = 168) Source of supply and continuation A final question sought to establish whether the usual source of pill supply after the first cycle was related to continuation. Given the high early dis- continuation rates of the sampled 1966-67 oral- contraceptive acceptor cohort, only a small portion of the discontinuers should have benefited from the April 1968 decision to allow distribution of conti- nuing supplies by field midwives. The reverse might be true for the minority who continued using oral contraceptives longer. Table 8 tends to confirm these expectations, but also shows that almost half of Table 8. Usual source of oral-contraceptive resupply for interviewed continuing users and discontinuing users taking more than one cycle % of continuing % of discontinuing Usual source of supply users users taking more (N =74) than one cycle (N = 129) clinic 46 73 midwife, at client's home 16 10 midwife, at midwife's home 4 6 midwife, at client's and midwife's home 7 1 clinic and midwife at client's home 18 8 clinic and midwife at midwife's home 8 2 private drug store and clinic 1 1 a Because of rounding to whole numbers, cumulative percent- ages do not equal 100. a Includes only acceptors taking more than I cycle of pills. b Rates based on less than 50 woman months of experience. c Rates significantly different (P <0.05). the continuing users still named the clinic alone as the usual source of supplies. Unless one assumes that these women preferred to return to the clinic at intervals for continuing supplies or that many field midwife posts were vacant for long periods, it appears that the field distribution policy was not implemented in all areas in 1968. On the other hand, about a quarter of the discontinuers had received new supplies from the midwives, suggesting the possiblity that field distribution was practised in some areas before the official policy change. Since most discontinuers stopped very soon after accep- tance, i.e., in 1966 or 1967, it was not surprising that almost three-quarters named the clinic only as the usual source of resupply. Very few acceptors used private sources ofsupply for oral contraceptives, presumably because the prices were 3-5 times those at Government clinics. Table 9 presents continuation rates for oral-contra- ceptive acceptors by usual source of resupply. At 6 months, oral-contraceptive acceptors availing them- selves of both midwife and clinic distribution were more likely to be still using oral contraceptives or another method. These rates were significantly higher than those characterizing acceptors relying on the clinic alone for resupply. Continuation rates for women relying mostly on the midwife for resupply were intermediate. This pattern holds through 12 months for first method rates, but changes for all-method continuation. Those women who relied mostly on the midwife for continuing supplies have the highest all-method continuation rate at 12 months. It can be seen in Table 9 that the women who relied on the midwife for resupply were much more likely . . 644 CEYLON: CONTRACEPTION IN A FIELD PROGRAMME 645 to be using a second method after discontinuing oral contraception than those discontinuers who relied chiefly on the clinic for continuing supplies (60% as against 27 %). DISCUSSION An important finding concerning study methodo- logy appeared in Table 5-namely, that acceptors who were sought out and interviewed at home were much less apt to be continuing the first or a later method than those previously interviewed at the health centre, hospital, or clinic. This finding points to a substantial upward bias in studies of the conti- nuation or use-effectiveness type that rely on volun- tary clinic returns for data, unless the proportion returning is very high. Although the possibility of bias has been widely speculated upon, to the best of our knowledge no data have been published suggesting its direction and degree. It seems likely that women lost to follow-up in this study had yet different continuation experiences from women inter- viewed at home or at the clinic, but possible bias here was reduced by the very high rate of completed interviews in the sample. The data presented in this study on age, living children, and open intervals of acceptors fall within the range of other national programme data (Ross et al., 1972). The same appears to be true for the data on previous contraceptive practice by acceptors, although the national data are not quite comparable since they characterize acceptors of other methods as well as the IUD and oral contraceptives. Although not derived from a national source, the Ceylonese IUD continuation rates are higher and those for oral-contraceptive acceptors lower than the observed international range. There are several possible reasons for lower oral contraceptive continuation rates in this study. Our life table calculations included the 7% ofwomen who purchased pills but never took them. It is not known how other reports have handled this problem, but, if such cases are excluded, continuation rates will be higher. In our judgment, these cases must be included to gain an accurate picture of overall programme performance. Since age is related to continuation, age-adjust- ment, if the Ceylon oral contraceptive sample were significantly younger than the national samples cited above, might narrow the observed differences. Furthermore, the higher estrogen content of oral contraceptives available in 1966 and 1967 compared with those available later makes it likely that side effects (and consequent discontinuation) will be more frequent in earlier than in later cohorts of acceptors. Program factors are important in deter- mining oral-contraceptive continuation rates, and these vary widely from country to country. Supply lines are difficult to maintain. Price is also a factor for some acceptors, and in Ceylon the pills were not free. It has been suggested on the basis of a demonstra- tion project in Thailand that a policy allowing the prescription and distribution of oral contraceptives by auxiliary personnel will both increase acceptance and improve continuation (Rosenfield, 1972). Clear- ly, the 1966-67 Ceylon cohort of oral-contraceptive acceptors did not benefit greatly from the 1968 policy decision to allow prescription and distribution by midwives. Table 7 suggested that distribution by mid- wives may have helped some women, but mostly those who continued on into 1968, and they might have continued in any case. Other Ceylonese data have failed to show that distribution by auxiliary personnel necessarily leads to higher continuation sales of oral contraceptives when applied to a national programme. A study of 242 1968-69 oral-contraceptive acceptors in Kuru- negala Province-an area culturally similar to Kalutara and Galle-yielded a 12-month first method continuation rate of 41, while in Batticaloa Province, a culturally dissimilar area, the 6-month continuation rate among 1969 oral-contraceptive acceptors was 30 (Wright, unpublished data, 1973). There was evi- dence that not all field midwives were actively implementing the new policy, however, and, in Batticaloa, many positions were vacant at the time of survey. Further, many women could not afford to buy the pills from month to month and the supply line was difficult to maintain. ln summary, while acceptance is likely to increase, it remains to be shown that a policy of prescription and distribution of oral contraceptives by auxiliary personnel can improve continuation rates outside a carefully controlled pilot area. Such a policy is amply justified, but is not likely to be effective in the absence of other critical program inputs (Nycander 1). The substantial gap between oral-contraceptive and IUD first method continuation rates in Table 4 1 Nycander, G. (1971) Family planning in the field. .. (Styrelsen for internationell utveckling SIDA mimeographed document). 646 NICHOLAS H. WRIGHT & TERRENCE PERERA would not be much reduced by adjusting for age, parity, and intention to space or limit. Continuation studies with larger samples attempting such adjust- ment show that differences, although narrowed, remain large. In the recently completed International Postpartum Program Follow-up Survey, first method IUD continuation rates were higher than those for oral-contraceptive acceptors, in 51 of 60 age-parity comparisons, a highly significant finding (Sivin, 1972). The all-method continuation rates in Table 4 show that discontinuing acceptors of oral contra- ceptives were more likely to shift to another method. At the end of a year, 30% of the discontinuing acceptors were using a second or later method, as compared with 20% of discontinuing IUD acceptors. On balance, the minority who changed and continued exchanged more effective methods for less effective ones like withdrawal or periodic abstinence. How- ever, the fact that so few changed after discontinuing the first method is discouraging and presents a serious challenge to the Ceylon programme, which offers sterilization, foam, and the condom, as well as the IUD and oral-contraceptives, and emphasizes continuing use of contraception, not simply the first method. All-method continuation rates for IUD acceptors in this study were comparable with those available from studies by Laing,' Sivin,2 and Tietze & Lewit (1971). All-method continuation rates for oral- contraceptive acceptors are more variable but not strikingly so. (More variation is seen in the U.S. studies by Tietze & Lewit, especially for oral- contraceptive acceptors.) These studies also showed "Laing, J. (1972) Use effectiveness offamily planning in the Philippines, 1970-1972 (University of the Philippines Population Institute mimeographed document). ' To be published as a Population Council Monograph. dramatic differences in extended use-effectiveness namely, pregnancy rates that, at 12 and 24 months, were 2-3 times higher for oral-contraceptive than for IUD acceptors. The differences were widest in our study, reflecting lower contraceptive continuation rates among acceptors of oral-contraceptives. The evidence thus suggests that IUD acceptors avoid unplanned pregnancy more effectively than do oral- contraceptive acceptors, and that the large increase in oral-contraceptive acceptance in 1968-69 did not imply as great an advance towards the demographic goals of the Ceylon programme as would have a comparable increase in IUD acceptance. This is obviously not to argue that oral contra- ceptives should be restricted or unavailable. There is evidence suggesting that their availability attracts women who would not otherwise be attracted to family planning programmes (Wright, 1972). Oral contraceptives, however, create difficulties in pro- gramme administration and impose a burden on acceptors, who must sustain motivation daily. Non- clinical prescription and distribution of pills, while necessary, is not sufficient to improve pill continua- tion. It does not solve the basic problems of main- taining supplies and following up acceptors. Initial method selection is critical. In a study carried out in the Philippines, Laing 3 has shown that continuation rates are more closely related to initial method selected than to other variables such as age and number of living children. Recognition of the prob- lems associated with oral contraceptives and of the fact that during their use, pregnancy rates are not inferior to those among IUD wearers, may enable clinic and field staff to provide more realistic coun- selling at the important time of method choice. 8 Laing, J. (1973) Differentials in Philippine contraceptive continuation and pregnancy rates (U.P. Population Institute mimeographed document). ACKNOWLEDGEMENTS The authors thank Mr A. Cooray, Health Educator, and Mr D. de Silva and Miss I. Sivapragasam, Research Assistants at the Family Planning Bureau, for their particularly active help in organizing the field work for the study; and Mr Dallas Fernando and Mr D. P. Guneratne, Department of Census and Statistics, and Mrs Joanna Pomeranz of the Population Council for their generous assistance in data processing. This study was made possible by the assistance of the Ford Foun- dation to the Ministry of Health, Govemment of Ceylon. CEYLON: CONTRACEPNTON IN A FIELD PROGRAMME 647 RtSUMI- CEYLAN: CONTINUATION DU RECOURS A LA CONTRACEPTION CHEZ DES UTILISATRICES DE CONTRACEPTIFS ORAUX OU DE DISPOSITIFS INTRA-UTERINS AU COURS D'UN PROGRAMME DE PLANIFICATION FAMILIALE Une enquete suivie a et faite parmi des utilisatrices de contraceptifs oraux ou de dispositifs intra-uterins pendant le programme de planification familiale en cours a Ceylan. Elle a fait ressortir d'importantes differences entre les deux methodes sous le rapport de la perseverance dans la pratique contraceptive et de l'efficacit6 dans la prevention des grossesses non desirees. Apres 2 ans, 30% des utilisatrices de contraceptifs oraux continuaient a employer la pilule, tandis que 17% d'entre elles avaient adopt6 une autre methode. Chez Jes femmes recourant aux dispositifs intra-uterins, les chiffres correspondants etaient de 68% et 7% respective- ment. Le taux global d'usage prolonge de la contraception apres 2 ans, essentiel pour le succes du programme, etait donc de 60% plus elev6 parmi les utilisatrices de dispositifs intra-uterins. La prise en consideration de differences entre les deux groupes, en relation notamment avec l'age, la parite et le niveau d'instruction, serait de nature a reduire mais non a supprimer 1'eart entre les taux de continuation selon la m6thode initiale choisie. Les taux de grossesse chez les utilisatrices de contra- ceptifs oraux ou de dispositifs intra-uterins refletent les differences entre les deux groupes concemant la perseve- rance dans la pratique contraceptive. Si l'on considere le nombre total de grossesses non d6sirees survenues apres 2 ans chez les femmes ayant utilise plusieurs methodes, on note un taux de grossesse de 44% chez les femmes ayant d6bute par les contraceptifs oraux et de 17% chez les femmes ayant utilise en premier lieu les dispositifs intra-uterins. De toute evidence, la decision prise en 1968 d'autoriser les sages-femmes a distribuer les contraceptifs oraux sous contr6le medical a amene un accroissement rapide du nombre des utilisatrices. Neanmoins, il ne semble pas qu'il en soit result6 une prolongation concomitante de la duree de l'usage de la methode. D'autres facteurs, comme le prix et les difficult6s d'approvisionnement, entrent en jeu. Ces donnees corroborent les resultats d'etudes faites en d'autres endroits. I1 apparait que l'utilisation des contraceptifs oraux pose davantage de problemes que celle des dispositifs intra-uterins, tout en ne temoignant pas d'une efficacite superieure dans la prevention des grossesses non desirees. Ce fait devrait etre present a l'esprit des conseillers en planification familiale lors du choix d'une methode de contraception. REFERENCES Rosenfield, A. G. (1972) Amer. J. Obstet. Gynec., 114, 942-949 Ross, J. A. et al. (1972) Reports on Population/Family Planning, No. 12 Sivin, I. (1972) Studies in Family Planning, 2, 248-256 Tietze, C. & Lewit, S. (1968) Demography, 5, 931-940 Tietze, C. & Lewit, S. (1971) Fertility and Sterility, 22, 508-513 Wright, N. H. (1968) Demography, 5, 745-756 Wright, N. H. (1972) Studies in Family Planning, 3, 305-308