Long-term contraceptive protection, discontinuation and switching behaviour Intrauterine device (IUD) use dynamics in 14 developing countries Marie Stopes International 1 Conway Street Fitzroy Square London W1T 6LP t: +44 (0)20 7636 6200 f: +44 (0)20 7034 2369 e: info@mariestopes.org w: www.mariestopes.org Registered charity number: 265543 Company number: 1102208 Marie Stopes International delivers quality family planning and reproductive healthcare to millions of the world’s poorest and most vulnerable women. Mohamed M. Ali, Rachael K. Sadler, John Cleland, Thoai D. Ngo and Iqbal H. Shah 02 (IUD) Intrauterine Authors andDevice acknowledgements Marie Stopes International This publication was produced by Marie Stopes International and the World Health Organization Marie Stopes International delivers quality family planning and reproductive healthcare to millions of the world’s poorest and most vulnerable women. Established in 1948, WHO works to make the enjoyment of the highest attainable standard of health a reality for every human being. Authors Mohamed M. Ali1, Rachael K. Sadler1, John Cleland2, Thoai D. Ngo3 and Iqbal H. Shah4 1.Evidence-Based Health Situation and Trends Assessment, Division of Health Systems and Services Development, Regional Office for the Eastern Mediterranean, World Health Organization, Cairo, Egypt. 2. Department of Population Studies, London School of Hygiene and Tropical Medicine, London, United Kingdom. 3. Research and Metrics, Health System Department, Marie Stopes International, London, United Kingdom. 4. Department of Reproductive Health and Research, World Health Organization, Geneva, Switzerland. The authors declare that they have no conflicts of interest. For citation purposes Ali, Mohamed M; Sadler, Rachael K; Cleland, John; Ngo, Thoai D and Shah, Iqbal H. Long-term contraceptive protection, discontinuation and switching behaviour: intrauterine device (IUD) use dynamics in 14 developing countries. London: World Health Organization and Marie Stopes International, 2011. Contents Marie Stopes International Contents 1. Executive summary 06 2. Introduction 08 3.Data, materials and methods 3.1 Data 3.2 Contraceptive calendar 3.3 Selection of IUD episodes 3.4 Reasons for discontinuation and switching 3.5 Covariates 3.6 Methods of analysis 12 12 12 12 12 12 13 4. Results 4.1 Background information for the 14 countries 4.2 Women’s characteristics 4.3 IUD discontinuation a. Reported failure b. Method-related reasons c. Probabilities of discontinuation by duration of use d. Differentials in discontinuation 4.4 Method-switching following method-related discontinuation a. Time to switching b. Differentials 4.5 Comparison of IUD and other modern methods 14 14 16 18 18 19 19 23 24 24 25 28 5. Discussion and conclusions 5.1 Data limitations 5.2 Main findings 30 30 31 6. Appendix 34 7. Glossary and References 48 Acknowledgements This report was written by Mohamed M. Ali, Rachael K. Sadler, John Cleland, Thoai D. Ngo and Iqbal H. Shah. All authors contributed equally to the conceptualising and writing of the report. Mohamed M. Ali and Rachael K. Sadler were in charge of data handling and analysis. The report was reviewed by the Department of Reproductive Health and Research of the World Health Organization (WHO). The publication of this report was made possible by support from the American people through the United States Agency for International Development (USAID). The contents are the responsibility of Marie Stopes International and the World Health Organization and do not necessarily reflect the views of USAID or the United States Government. The authors would like to thank Catherine d’Arcangues of WHO’s Department of Reproductive Health and Research and the following individuals from Marie Stopes International (MSI) for their critical review of the report: Tania Boler, Louise Lee-Jones, Nicole Gray, Geraldine Ellis, Chris Duncan and Dana Hovig. Thanks are also due to Vicky Anning, who edited the report. Cover photograph © Marie Stopes International Support for this research was provided by Marie Stopes International (MSI). 03 04 Intrauterine Device (IUD) Marie Stopes International Intrauterine Device (IUD) Marie Stopes International Tables and figures Figure 1: Percentage of contraceptive users by types of contraceptive method used, by year 08 Table 1: P ercentage using any method of contraception that are using the IUD among women who are married or in union aged 15-49 and those using IUD among the users of any method of contraception, 2009 Figure 2a: IUD all-cause discontinuation probabilities at 12 months by country 20 10 Figure 2b: IUD all-cause discontinuation probabilities at 24 months by country 20 Table 2: Country background information 14 Figure 2c: IUD all-cause discontinuation probabilities at 36 months by country 20 Table 3: All contraceptive methods and IUD episodes reported in the calendar, by surveys 16 Figure 3a: IUD discontinuation probabilities due to method dissatisfaction at 12 months by country 21 Figure 3b: IUD discontinuation probabilities due to method dissatisfaction at 24 months by country 21 Table 4: P ercentage of IUD and other modern method (MM) episodes contributed by women of specified characteristics 17 Figure 3c: IUD discontinuation probabilities due to method dissatisfaction at 36 months by country 21 Table 5: Reasons for IUD discontinuation, by country 18 Figure 4a: IUD method-related discontinuation probabilities at 12 months, by residence 22 Table 6: Median cumulative cause-specific discontinuation probabilities per 100 episodes 19 Figure 4b: IUD method-related discontinuation probabilities at 12 months, by education 22 Table 7: Median conditional cause-specific discontinuation probabilities per 100 episodes 19 Figure 4c: IUD method-related discontinuation probabilities at 12 months, by motivation for use 23 Table 8: Status at three months after discontinuing IUD for method-related reasons 24 Figure 4d: IUD method-related discontinuation probabilities at 12 months, by wealth 23 Table 9: Duration of modern methods and IUD episodes 28 Figure 5: Switching behaviours three months post IUD discontinuation 25 Figure 6a: Switching to a modern FP method within three months of IUD discontinuation, by residence 26 Box 1: Facts about the intrauterine device (IUD) 08 Figure 6b: Switching to a modern FP method within three months of IUD discontinuation, by education 26 Figure 6c: Switching to a modern FP method within three months of IUD discontinuation, by motivation for use 27 Table A.1: Cumulative cause-specific discontinuation probabilities (per 100 episodes) 34 Figure 6d: Switching to a modern FP method within three months of IUD discontinuation, by wealth 27 Table A.2: Conditional 12-month cause-specific discontinuation probabilities (per 100 episodes) 36 Figure 7: All-cause discontinuation rates at 12 months for the four main modern, reversible methods 29 Figure 8: Discontinuation rates at 12 months due to failure of the four main modern, reversible methods 29 Table A.3: M ethod-related discontinuation probabilities (per 100 episodes), disaggregated by selected characteristics 36 Table A.4: Status at three months after discontinuing IUD for method-related reasons 38 Table A.5: Status at three months after discontinuing IUD for method-related reasons: by residence 40 Table A.6: Status at three months after discontinuing IUD for method-related reasons: by education 42 Table A.7: Status at three months after discontinuing IUD for method-related reasons: by motivation for use 44 Table A.8: Status at three months after discontinuing IUD for method-related reasons: by wealth status 46 Figure 9: P ercentage of women who switched to any modern method within three months post discontinuation due to method-related reasons, by method 30 05 06 Executive summary Marie Stopes International Executive summary Marie Stopes International 1. Executive summary The intrauterine device (IUD) is one of the most widely used, reversible contraceptive methods in the world. Women living in China constitute about 70% of the 106 million global users. Modern IUDs are a safe, highly effective and cost-effective method of contraceptive. However, IUD uptake varies significantly across the globe. Some countries have wide usage and others have very little or no usage. For instance, use among all married women of reproductive age is highest in China (where 33% of all married women use IUDs, as well as the Scandinavian countries (18%), Asian nations (13%) and the Near East and North Africa (12%). Approximately seven percent of all married European and Russian women and about five percent of women from Australia and New Zealand rely on the IUD as a method of contraception. Overall IUD use is lowest in Northern America (1.5%) and sub-Saharan Africa (0.8%). These figures reveal that IUDs remain underutilised, especially in Northern America, Oceania, South Asia and sub-Saharan Africa, in spite of the IUD’s many benefits. The aim of this report is to provide an in-depth assessment of IUD use-dynamics in developing countries, with a focus on discontinuation and method-switching, and to compare users of IUDs with users of other methods in these regards. Contraceptive calendar data from 14 nationally representative surveys were used for the analysis. All but two of the surveys were conducted between 1998 and 2008. All major developing regions are represented with the exception of sub-Saharan Africa, for which no suitable survey data were available. Most analyses focused on contraceptive-use episodes, with particular attention to length of use, reasons for stopping and method-switching following stopping. Even within the 14 study countries, which were selected for reasonably high overall contraceptive use and a minimum of 300 IUD use-episodes, the contribution of the IUD to contraceptive protection is immensely variable. It accounts for about two percent of users in some countries but over half of users in other countries. This variability probably reflects policy choices about which methods to promote and preference biases in family planning services. In view of the advantages of the IUD over alternative reversible methods in terms of its low failure rate and high continuation, the virtual absence of IUDs from the overall method mix in many countries is deeply unfounded. While the IUD, along with other long-acting methods such as implants, is well suited to the needs of couples who have all the children that they desire, about half of all IUD use appears to be motivated by the wish to space or postpone childbearing. IUD use is compatible with breastfeeding and thus has a potentially valuable role in postpartum contraceptive protection. In terms of urban-rural residence and household wealth, IUD users differ little from users of other modern methods. In five of the 14 countries, IUD users were less well educated than other method users. However, in the majority of countries, this difference was negligible or reversed. © Marie Stopes International / Peter Barker The main reasons for stopping IUD use imply dissatisfaction with the method. However, more intensive studies are needed to gain more insight into precise concerns. The probability of abandoning IUDs for method-related reasons is approximately constant over the three years following adoption, though this pattern varies between countries. Huge inter-country variations in the probability of stopping IUD use were observed. After 24 months more than 40% of women had stopped use in three countries compared with 20% or less in another three countries. The reasons for variation in the willingness or ability of women to persist with the method are unclear. In contrast to pronounced inter-country differences, discontinuation does not vary appreciably by women’s schooling, household wealth or urbanresidence. Thus the method is equally suitable for all socio-economic and residential strata. Thus, an IUD offers, on average, three times the length of contraceptive protection offered by other modern reversible methods. © Marie Stopes International / Peter Caton Following IUD discontinuation for method-related reasons, half of all women on average switched to another reversible modern method within three months and an additional 11.5% switched to a traditional method. In most countries the pill was the most common destination method and voluntary sterilisation was rare except in three Latin American countries. The percentage who switched methods ranges from 40% to 80% across the 14 countries, indicating very great variability in access to, or acceptability of, alternatives to the IUD. Switching to another method following discontinuation is more common for those who stopped IUD use than those who stopped use of other modern methods. The analysis confirms the huge advantages of the IUD over alternative (mainly hormonal) spacing methods in terms of reported failure and discontinuation in all 14 countries. Whereas more than 40% of users of the pill, injectable methods and condoms stopped use within 12 months, the equivalent figure for IUDs was only 13%. Similarly, while the median length of uninterrupted use of IUDs is typically 30 or more months, the median for other methods is typically ten months. Thus an IUD offers, on average, three times the length of contraceptive protection offered by other modern reversible methods. 07 08 Introduction Marie Stopes International In sub-Saharan Africa, the use of the IUD is less common. Only two percent of users rely on the IUD. Introduction Marie Stopes International 2. Introduction Intrauterine devices (IUDs) for preventing pregnancy were first introduced more than a century ago. In 1909, the first paper reporting the use of a ring made of silkworm gut was published1. Advances in contraceptive technologies over the last 35 years have resulted in the availability of copperbearing and levonorgestrel-releasing IUDs, which are easier to insert and remove. Safety and effectiveness vary by type of IUD. Copper T380A IUD was found to be the most effective copper-containing device 2. Among IUD users, contraceptive failure is below one percent by the end of the first year 3. The use of IUDs has increased over the past 25 years and it is now the most widely used reversible contraceptive method globally (Figure 14-5). In 2007, 22.6% of 721 million users of any method of contraception worldwide relied on IUDs to prevent pregnancy 5. Among all women aged 15-49 who were married or in union in 2007, female sterilisation was used by 20.3% and IUDs by 14.2%. Thus, the IUD is the leading reversible method and the second most commonly used method when all reversible and irreversible methods are considered. % World Developing 90 Injectable 80 IUD 70 Condom Withdrawal/Rhythm Other The Fourth Edition of Medical Eligibility Criteria for Contraceptive Use (MEC) 6 considered more than 100 personal characteristics and reproductive history aspects to examine the suitability of IUDs for each one. 60 IUDs can be safely and effectively used by women under most conditions. However, IUDs do not protect against sexually transmitted infections (STI) or the human immunodeficiency virus (HIV). MEC recommended the correct and consistent use of condoms, either alone or with another contraceptive method, if there is a risk of STI/HIV, including during pregnancy or post-partum. 50 40 30 20 10 1983 2007 1983 2007 1983 The figure of 16% for Africa is influenced by the relative dominance of IUDs in Northern Africa, where 37% of all users rely on the IUD. In Egypt and Tunisia, one in two contraceptive users has an IUD. In sub-Saharan Africa, the use of the IUD is less common. Only two percent of users rely on the IUD. Medical eligibility criteria for IUDs Developed Male sterilisation Pill Both the global figure and the figure for Asia reflect the predominance of the IUD in China, where this method accounts for one in two contraceptive users. Excluding China from the global estimates of users of any method and IUD users, IUDs account for only 12% of all use, halving the percentage of users of IUDs among all users worldwide. BOx 1: Facts about the intrauterine device (IUD) FIGURE 1: Percentage of contraceptive users by types of contraceptive method used, by year Female sterilisation The overall worldwide trends and levels mask the variation in use of any contraceptive method, and of IUD use by region and sub-region, as well as by country and within countries. The latest update of World Contraceptive Use 20095 indicates that the prevalence of IUD use is highest in Asia, where one in four users of any method of contraception relies on the IUD, followed by Europe with about one IUD user among five (Table 1). 2007 Source: United Nations, 2009 BOx 1: Facts about the intrauterine device (IUD) What is an IUD? The letters IUD stand for ‘intrauterine device’. IUDs are small, T-shaped devices made of flexible plastic. A health-care provider inserts an IUD into a woman’s uterus to prevent pregnancy. There are two popular brands of IUD available globally – ParaGard and Mirena. The ParaGard IUD contains copper and is effective for 12 years. The Mirena IUD releases a small amount of the hormone progestin and is effective for five years. It also recommended that copper-containing IUD (Cu-IUD) or levonorgestrel-releasing IUD (20 μg per 24 hours) or both should not be used under the following conditions: a.during pregnancy; b.immediately after septic abortion; c. post-partum puerperal sepsis; d.before evaluating unexplained vaginal bleeding; e.persistently elevated β-human chorionic gonadotropin (β-HCG) levels or malignant disease; f.cervical cancer (while awaiting treatment); g.current breast cancer (for initiating levonorgestrel IUD (LNG-IUD)); h.endometrial cancer; i.uterine fibroids with distortion of the uterine cavity; j.distorted uterine cavity (any congenital or acquired uterine abnormality distorting the uterine cavity in a manner that is incompatible with IUD insertion); k.current pelvic inflammatory disease (PID); l.current purulent cervicitis or chlamydial infection or gonorrhoea; and m.pelvic tuberculosis. 09 10 Introduction Marie Stopes International Introduction Marie Stopes International In spite of this method’s attractive features, IUDs remain underutilised, especially in Northern America, Oceania, South Asia and sub-Saharan Africa. IUD use exceeds 30% among all women who are married or in union in just 11 countries (China, Cuba, Estonia, Egypt, Kazakhstan, Kyrgyzstan, Mongolia, Tajikistan, Turkmenistan, Uzbekistan and Viet Nam) 5. As contraceptive prevalence in a population increases, success in avoiding unwanted pregnancy depends less on initial contraceptive uptake and more on effective and persistent use. Once a method is adopted, it is important that couples have quick and easy access to other methods, should the initial one be deemed unsuitable. Contraceptive method use dynamics therefore becomes an important determinant of success in achieving reproductive intentions. In spite of this method’s attractive features, IUDs remain underutilised, especially in Northern America, Oceania, South Asia and sub-Saharan Africa. This report aims to provide detailed information on the dynamics of IUD use in developing countries, using data from the most recent Demographic and Health Surveys (DHSs). More specifically, it provides the socio-demographic profile of IUD users, continuation of use and reasons for the discontinuation of use at 12, 24 and 36 months. In addition, it provides information on method-switching following discontinuation for method-related reasons. The ultimate aim of this report is to provide information for policies and programmes to strengthen services for improving the continuation of IUD use and shortening the time for switching to a modern method for women who do not want another child. TABLE 1: Percentage using any method of contraception that are using the IUD among women who are married or in union aged 15-49 and those using IUD among the users of any method of contraception, 2009 Region % using any method % using IUD % of all users relying on IUD World 62.9 14.2 22.6 More developed 69.7 9.1 13.1 Europe 70.5 13.6 19.3 Northern America 72.9 1.7 2.3 Less developed 61.7 15.1 24.5 Africa 28 4.5 16.1 Asia 67 17.7 26.4 Latin America & the Caribbean 71.7 7.2 10 Oceania 58.6 1 1.7 Source: United Nations, 2009 © Marie Stopes International 11 12 Data, materials and methods Marie Stopes International Data, materials and methods Marie Stopes International 3. Data, materials and methods 3.1 Data 3.4 Reasons for discontinuation and switching 3.6 Methods of analysis The data used in this analysis come from 14 Demographic and Health Surveys (DHSs) conducted between 1993 and 2008. The main criterion for inclusion in this report is the availability of at least 300 IUD episodes that allows disaggregated analysis. Five of the surveys come from Latin America & the Caribbean, five from North Africa/ West Asia and four from South-Central/South-East Asia. The lack of surveys in sub-Saharan Africa reflects the generally low contraceptive prevalence rate and particularly the low use of IUDs (Table 1). The reasons given for discontinuation were grouped into the following four categories: •reported method failure (i.e. the respondent became pregnant while using the IUD) • a desire to become pregnant •no further need (i.e. sexual abstinence due to illness or marital dissolution) • method-related reasons. Reasons for discontinuation were analysed using single-decrement life-table methods and reported per 100 episodes of use. These rates give hypothetical cause-specific probabilities of discontinuing use for users who do not stop for any other reason. Life-table methods were also used to calculate the median duration of use. Cumulative cause-specific discontinuation rates at 12, 24 and 36 months of use are presented. It is also of interest to investigate the probability of discontinuing with the IUD due to method-related reasons by time elapsed since starting the method. The analysis investigated the probability of discontinuing between 12 and 24 months and between 24 and 36 months. 3.2 Contraceptive calendar As of 1990, the DHS programme has included a contraceptive calendar for countries where contraceptive prevalence is considered to be relatively high (≥40%). During the interview, previous births and current pregnancies are entered into a grid spanning five years preceding the survey. Abortions are also entered into the calendar, although it is not possible to distinguish between spontaneous and induced abortions. Around these reproductive events, dates of contraceptive use are entered and reasons for discontinuation are noted. Calendars of contraceptive use have been found to generate more complete and accurate data on past contraceptive use than other formats7-9. 3.3 Selection of IUD episodes The units of analysis in this study are episodes of IUD use and the time to first event following discontinuation of IUD for method-related reasons. An episode is defined as a period of uninterrupted use (in months) that may or may not have ended. A switch to a new method (or a break between methods) indicates the start of a new episode. The episodes analysed in this report were contributed by women who were either married or in union at the time of use. Calendar data for the three months prior to the interview data were omitted to avoid the problem of under-reporting of first trimester pregnancies10, as including these would underestimate the reported IUD failure rate. This final category includes reasons that imply some degree of dissatisfaction with the IUD, such as side effects, health concerns, medical advice, access and availability, desire to switch to a permanent method, inconvenience of use and cost. While women may discontinue with their IUD for multiple reasons, the survey only recorded the main reason. For the purposes of this report, the categories for discontinuation of interest were ‘reported IUD failure’ and ‘method-related reasons’, as other reasons are driven by user desires and are not related to the method itself. 3.5 Covariates Four covariates are used in this analysis. Women’s usual place of residence is either rural or urban. Education level was re-grouped into two: primary schooling or less, and secondary schooling or higher. Motivation for use (i.e. using the IUD to limit births or to space them) was derived by comparing the total number of children desired and number of living children at the start of the episode. If total desired size is less than the number of living children, the motive for use is inferred to be limitation. If desired size equals or exceeds actual size, the motive is classified as spacing. The DHS wealth quintiles, based on household possessions and characteristics, were re-coded into tertiles, with those in the top third (the richer) compared with those in the bottom third (the poorer). Multiple-decrement life-table methods were employed to estimate the status of women three months after having discontinued the IUD due to method-related reasons. It was assumed that these women would still require contraception as they had not stopped due to a desire to become pregnant but because of dissatisfaction with the IUD. The three month mark was chosen because few women switch to another method after this length of time. All analyses were done using STATA 10.1 (StataCorp. 2009. College Station, TX: StataCorp LP) and appropriate adjustments for the survey weights and clusters were made. For the purposes of this report, the categories for discontinuation of interest were ‘reported IUD failure’ and ‘method-related reasons’. © Marie Stopes International 13 14 Results Marie Stopes International 4. Results Results Marie Stopes International TABLE 2: Country background information Country (abbreviation) Date of fieldwork Number of women 4.1 Background information for the 14 countries As can be seen in Table 2, current use of any contraceptive method among currently married women in the 14 study countries ranges from 44.6% in Bangladesh to 78.5% in Viet Nam, while current use of modern methods ranges from 17.7% in Bolivia to 68.2% in Colombia. The use (popularity) of the IUD varies from 2.2% in both Bangladesh and the Dominican Republic to 42% in Kazakhstan. While current use of modern methods is high in the Dominican Republic (65.8%), use of the IUD is low (2.2%). Conversely, the level of modern contraceptive prevalence in Kazakhstan (52.7%) is mainly a result of IUD use (42%). Median use of IUD is 8.6% across the 14 countries studied. In addition to contraceptive use, Table 2 presents the total fertility rate (TFR) and unmet need for family planning (contraception). TFR and unmet need ranges from 1.9 to 4.8% in Viet Nam and 4.8 to 24.3% in Bolivia, respectively. Levels of TFR and unmet need are significantly correlated with current use of any method and modern method (TFR: -0.78 and -0.64, respectively and unmet need: -0.91 and -0.79). % currently using:* Any method Modern method IUD TFR (15-49) Unmet need % Latin America & the Caribbean Bolivia (BO) 1993/1994 8,603 45.3 17.7 8.1 4.80 24.3 Colombia (CO) 2004/2005 41,344 78.2 68.2 11.2 2.40 5.8 Dominican Republic (DR) 2002 23,384 69.8 65.8 2.2 3 10.9 Nicaragua (NC) 1997/1998 13,634 60.3 57.4 9.1 3.90 14.7 Peru (1) (PE) 2003/2005 11,717 71.3 47.6 5.6 2.60 8.1 Egypt (EG) 2008 16,527 60.3 57.6 36.1 3 9.2 Jordan (JO) 2007 10,876 57.1 41.9 22.3 3.60 11.9 Morocco (MA) 2003/2004 16,798 63 54.8 5.4 2.50 10 Turkey (TR) 1998 8,576 63.9 37.7 19.8 2.61 10.1 Bangladesh (BD) 1993/1994 9,640 44.6 36.2 2.2 3.44 19.4 Indonesia (ID) 2007 32,895 61.4 57.4 4.9 2.60 9.1 Kazakhstan (KK) 1999 4,800 66.1 52.7 42 2.05 8.7 Philippines (PH) 2003 13,633 48.9 33.4 4.1 3.50 17.3 Viet Nam ( VN) 2002 5,665 78.5 56.7 37.7 1.87 4.8 12,675 62.2 53.8 8.6 2.81 10.1 Northern Africa/Western Asia South central/South-eastern Asia Total (Median) *among currently-married women aged 15-49, for Bangladesh (10-49) (1) Peru is a continuous survey © Marie Stopes International 15 16 Results Marie Stopes International Results Marie Stopes International 4.2 Women’s characteristics Table 3 shows the total number of women who contributed at least one episode and number of episodes used of any method. It also shows those who reported IUD episodes during the five-year calendar, as well as the mean number of episodes reported, by survey. Overall 137,884 contraceptive episodes started during the five-year period of the calendar, 18,484 (13.4%) of which were IUD episodes. The mean number of IUD episodes per woman was slightly more than one (1.05), which was lower than the mean number of all contraceptive method episodes (1.61). Table 4 compares the characteristics of IUD users with those of users of other modern contraceptive methods. In terms of urban-rural residence, on average across all 14 countries, there is little difference between IUD users and users of other modern methods: 71% of IUD episodes were recorded in urban women compared with 67% for other methods. Differences were observed, however, in Peru, Nicaragua and Indonesia, where more IUD episodes were contributed by women in urban settings than was the case for users of other modern methods. The opposite was found in Kazakhstan, Viet Nam and the Philippines. Nearly one-third (30%) of the IUD episodes in the 14 surveys combined were contributed by women educated to at least secondary school level. The corresponding figure for other modern methods was higher (42%) and this pattern was observed in the majority of the 14 countries. However, this difference was pronounced in only five countries (Dominican Republic, Egypt, Indonesia, Nicaragua and Peru). In terms of motivation for IUD use, women appeared to use the IUD equally for both spacing and for limiting births, with each motivational factor contributing 50% of the episodes. In all countries, women were more likely to use modern methods other than IUDs to space births (a median of 59%), with the exception of Egypt. Around 40% of IUD episodes were contributed by women from more affluent households, and the same finding was observed for women using other modern methods. However, there were mixed findings within the countries. Wealthier women in Viet Nam and Kazakhstan contributed fewer IUD episodes than episodes of other modern methods, whereas the opposite was found in Indonesia, Peru and Egypt. © Marie Stopes International / Susan Schulman TABLE 4: Percentage of IUD and other modern method (MM) episodes contributed by women of specified characteristics TABLE 3: All contraceptive methods and IUD episodes reported in the calendar, by surveys Country All methods Women episodes IUD Women n mean episodes n mean Bangladesh 4,253 6,806 1.60 358 369 1.03 Bolivia 2,797 4,435 1.59 464 495 1.07 Colombia 18,127 30,415 1.68 2,768 2,843 1.03 Dominican Republic 8,896 14,879 1.67 652 684 1.05 Egypt 7,346 10,024 1.36 4,394 5,008 1.14 Indonesia 14,668 20,341 1.39 663 682 1.03 Jordan 6,079 10,270 1.69 2,381 2,624 1.10 Kazakhstan 1,856 3,253 1.75 1,030 1,125 1.09 Morocco 5,428 10,126 1.87 490 506 1.03 Nicaragua 5,122 7,866 1.54 1,228 1,291 1.05 Peru 2,671 5,186 1.94 248 267 1.07 Philippines 3,764 5,783 1.54 298 304 1.02 Turkey 3,034 4,909 1.62 1,007 1,046 1.04 Viet Nam 2,451 3,551 1.45 1,170 1,241 1.06 Median (4,688) (7,336) (1.61) (835) (865) (1.05) Country % Urban MM IUD % Secondary + MM IUD % Spacing MM IUD % Upper third (wealth) MM IUD Bangladesh 17 17 73 72 42 36 46 41 Bolivia 90 87 26 21 58 51 65 66 Colombia 80 78 22 26 67 51 48 42 Dominican Republic 70 76 44 30 77 73 48 59 Egypt 36 43 42 31 50 52 22 42 Indonesia 42 67 44 19 64 49 37 71 Jordan 86 87 7 7 60 57 41 47 Kazakhstan 70 53 n/a n/a 77 66 41 25 Morocco 59 69 77 72 58 47 37 46 Nicaragua 64 77 55 41 63 58 43 45 Peru 77 90 21 10 60 48 41 64 Philippines 51 40 24 28 55 44 24 17 Turkey 73 73 67 71 53 40 47 41 Viet Nam 25 13 32 34 42 40 48 22 Total (median) (67) (71) (42) (30) (59) (50) (42) (43) MM= Modern methods (include pill, injections, vaginal methods, condom and Norplant) 17 18 Results Marie Stopes International Results Marie Stopes International 4.3 IUD discontinuation Table 5 shows the number of IUD episodes and the percentage discontinued by reason across the 14 countries during the five-year period. On average for the 14 study countries, 66% of IUD episodes were still continuing at the time of the survey. The dominant reason for stopping was health concerns or side effects (15.7%); 2.5% episodes were discontinued because of reported method failure, 5.9% due to a desire for a child and 3.5% because of no further need. Bangladeshi women showed the highest levels of discontinuation due to health concerns at 39.8% with Indonesia showing the lowest at 8.2%. Table 6 presents the medians of the country-specific, cumulative overall and cause-specific discontinuation rates at 12, 24 and 36 months, per 100 episodes. (See appendix Table A1 for country-specific rates.) Across all 14 countries, the median probability of discontinuing an IUD for all reasons was 13.2% at 12 months, 27.5% at 24 months and 41.5% at 36 months. b. Method-related reasons. Across all countries and at all time periods, the main reasons for discontinuation were related to the method itself; the median methodrelated discontinuation rates were 8.9% at 12 months, 16.7% at 24 months and 27% at 36 months. The highest method-related discontinuation rate at 12 months was observed in Bangladesh (32.2%), while rates of around 18% were found in the Dominican Republic, Peru and Nicaragua (Figures 3a, 3b and 3c). At 36 months, over one-third of all episodes in the Dominican Republic (46.5%), Nicaragua (40.9%) and Peru (34.4%) had been discontinued due to dissatisfaction with the method, and over half in Bangladesh (53.6%). In terms of inter-country variability, three countries showed consistently low probabilities of IUD discontinuation at all time periods, namely Indonesia, Turkey and Kazakhstan, while Bangladesh and the Dominican Republic showed the highest (Figures 2a, 2b and 2c, please see overleaf ). a. Reported failure. Discontinuation due to reported method failure (i.e. the user became pregnant while using the IUD) was low across the 14 countries. Just over one per 100 episodes (1.3) failed in the first year, rising slightly to nearly four out of 100 at three years of use (Table A1). At 12 months of use, the reported failure rate was lowest in Bangladesh (0.4%) and highest in Kazakhstan (3.2%). At 24 months, the median reported failure rate was 2.3%, with the lowest rate observed in the Philippines (0.6%) and the highest in Kazakhstan (5.8%). By 36 months, the median overall reported failure rate was 3.9%. The lowest rate at this time point was found again to be in the Philippines (0.6%) and the highest in Colombia (6.9%). In terms of inter-country variability, three countries showed consistently low probabilities of IUD discontinuation at all time periods, namely Indonesia, Turkey and Kazakhstan. TABLE 5. Reasons for IUD discontinuation, by country Country Total IUD episodes Currently using (%) c. Probabilities of discontinuation by duration of use. Table 7 shows overall and cause-specific probabilities of discontinuing the IUD between 12 and 24 months (denoted in the table as 24 | 12) and between 24 and 36 months (shown as 36 | 24). (See appendix Table A2 for country-specific findings.) The median reported failure rate remained low at 1.4% in the second year of use, with little discernible increase in the third year (1.5%). While the reported failure rate in the majority of countries remained either stable or decreased, the rates in both Indonesia and Jordan doubled in the third year of use compared with the second year (1.3% to 3% and 1.6% to 3.1% respectively). TABLE 6. Median cumulative cause-specific discontinuation probabilities per 100 episodes. (Median values for the 14 countries) Discontinued (%): Method failure Health concens Other reasons Want a child No further need Month All reasons Failure Methodrelated Want a child No further need 12 13.2 1.3 8.9 1.4 1.6 Bangladesh 369 44.5 0.8 39.8 2.4 5.2 7.3 Bolivia 495 71.9 1.4 15.6 1 5.7 4.4 Colombia 2,843 68.2 4.3 20 2.2 3.4 1.9 24 27.5 2.3 16.7 5.2 2.8 Dominican Republic 684 47.8 2.8 27.9 5.1 8.3 7.9 36 41.5 3.9 27 11.1 4.9 Egypt 5,008 63.3 2.1 11.6 0.4 19.6 3 Indonesia 682 77.8 2.8 8.2 1.9 5.6 3.8 Jordan 2,624 63.6 2.8 13.5 1.4 17.3 1.5 Kazakhstan 1,125 73 5.2 12 1.2 6.8 2 Morocco 506 60.7 1.6 22.5 1 11.1 3.2 Nicaragua 1,291 52.4 3.3 26.6 4.3 6.2 7.4 Peru 267 63.8 2.3 26.3 0.4 5.6 1.5 Philippines 304 73.1 0.7 15.8 3.9 2 4.6 Turkey 1,046 72.9 2.1 12.9 0.6 5.4 6.1 Viet Nam 1,241 75.2 3.1 12.7 0.6 6.8 1.5 Total (median) 865 66 2.5 15.7 1.3 5.9 3.5 TABLE 7. Median conditional cause-specific discontinuation probabilities per 100 episodes. (Median values for the 14 countries) Month All reasons Failure Methodrelated Want a child No further need 12 | 24 15.6 1.4 8.5 3.4 1.1 36 | 24 17.5 1.5 8 5.4 1.7 19 Results Marie Stopes International Results Marie Stopes International Figure 2a: IUD all-cause discontinuation probabilities at 12 months by country Figure 3a: IUD discontinuation probabilities due to method dissatisfaction at 12 months by country Percent 37.3 Figure 2B: IUD all-cause discontinuation probabilities at 24 months by country 19.7 Dominican Republic Colombia Nicaragua 12.7 Peru 11.4 Morocco Viet Nam 6.8 Philippines 6.4 8.9 18 Percent 46.5 53.9 12.6 12.7 20 20.2 Bangladesh 11.7 17.5 32.6 Dominican Republic Bangladesh Dominican Republic Nicaragua Egypt Peru Morocco Colombia Jordan Bolivia Philippines Kazakhstan Viet Nam Turkey Indonesia Figure 2C: IUD all-cause discontinuation probabilities at 36 months by country 10.1 15.9 Morocco 10 10 8.3 14 Jordan 20 29.3 Nicaragua 30 33.5 Colombia 32.6 Egypt 23.2 29.7 Viet Nam 22.3 27.7 Kazakhstan 20.4 27.3 Indonesia 15.3 17.7 25.7 Turkey 30 27.8 Philippines 40 40 Peru 44 48.5 Bolivia 50 Figure 3C: IUD discontinuation probabilities due to method dissatisfaction at 36 months by country Percent Percent Bangladesh Dominican Republic 20.3 Nicaragua 18.4 28.4 Colombia 16.4 27.2 Philippines 10.9 16.1 27.1 Morocco Bangladesh Dominican Republic Nicaragua Egypt Jordan Peru Morocco Bolivia Colombia Philippines Turkey 10 15.4 26.9 Bolivia 20 Jordan 30 24.6 Viet Nam 34.4 Egypt 40 40.9 40 Viet Nam 38.8 51.2 46.5 Turkey 33.3 53.6 50 Kazakhstan 30.8 48.2 66 Indonesia 30.5 30.6 43 47.1 63.1 Peru 58.1 Kazakhstan 60 50 40 30 20 10 6.3 8.8 17.5 Figure 3B: IUD discontinuation probabilities due to method dissatisfaction at 24 months by country Percent 20 6 8.1 Bolivia Bangladesh Dominican Republic Nicaragua Peru Colombia Morocco 10 7.5 Jordan 20 Indonesia 20.1 Turkey 15.5 16.7 28.2 Egypt 13.8 Philippines 12.5 Viet Nam 12 Jordan 11.9 Bolivia 11.8 Egypt 11.4 Kazakhstan 10.4 Indonesia 10 9.6 Turkey 20 24.2 32.2 30 Kazakhstan 30 Bangladesh Percent Indonesia 20 21 22 Results Marie Stopes International Results Marie Stopes International Across all 14 countries, the probability of stopping for method-related reasons declines over time. In the first 12 months of use, 8.9% of episodes were terminated for this reason. For women who continued with use for 12 months or more, the probability of discontinuing in the second year fell to 8.5%. For those who completed 24 months or more of use, it fell further to 8% in the third year. However, examination of country-specific results reveals a variety of patterns. In some countries (e.g. Egypt and Jordan) discontinuation was rather constant over the three years. In Indonesia and Viet Nam, the rate was relatively high in the first 12 months but low in the second and third year. In other countries (e.g. Nicaragua and Turkey), it was high in the first year, fell in year two but rose in the third year. Figure 4a shows that IUD discontinuation was marginally higher on average among rural than urban women, but the country-specific results in Table A3 indicate a fair number of exceptions (notably Philippines and Viet Nam) where discontinuation among urban women was about twice as high as among rural women. d. Differentials in discontinuation. It is of interest to know whether or not women with different characteristics also vary in their propensity to discontinue IUD use for method-related reasons. This topic is addressed in Figures 4a, 4b, 4c and 4d. Results are summarised by box and whisker plots for discontinuation at 12 months. The height of each box indicates the range within which half of estimates for the 14 countries fall and the horizontal line in the box shows the median value. The vertical lines protruding from each box (the whiskers) show the predicted range within which 90% of national estimates fall. Finally, dots lying beyond the end of the whiskers represent outliers. In this analysis, all outliers represent Bangladesh. As expected, motivation for use (Figure 4c) influenced discontinuation, with a higher 12 month risk of stopping among spacers than limiters (11.4% versus 7.9%). This contrast was most pronounced in Peru (25.2% versus 9.9%) but, conversely, was negligible or slightly reversed in Bolivia, Dominican Republic and Morocco. The average differences by women’s education were small; the overall median was slightly lower for those with secondary or higher schooling than for the less well educated, but the two boxes in Figure 4b shows a great degree of overlap. The big exception is Indonesia, where the probability of stopping was ten times greater among better educated women than among the less well educated. Differences according to household wealth tended to be small (Figure 4d). © Marie Stopes International / Peter Caton FIGURE 4A: IUD method-related discontinuation probabilities at 12 months, by residence Discontinuation rate per 100 episodes FIGURE 4B: IUD method-related discontinuation probabilities at 12 months, by education Discontinuation rate per 100 episodes FIGURE 4C: IUD method-related discontinuation probabilities at 12 months, by motivation for use Discontinuation rate per 100 episodes FIGURE 4D: IUD method-related discontinuation probabilities at 12 months, by wealth Discontinuation rate per 100 episodes 30 30 30 30 20 20 20 20 10 10 10 10 rural urban <primary secondary+ spacer limiter poor rich 23 24 Results Marie Stopes International Results Marie Stopes International 4.4 Method-switching following method-related discontinuation Table 8 displays contraceptive status at three months following discontinuation. Across all 14 countries the median percentage of women who were already pregnant was 12.4% three months after discontinuation. The highest observed value in the analysis was in Bolivia (23.7%), while the lowest was observed in Morocco (2.8%). It is likely that most of these pregnancies were unintended. A quarter of women were still at risk of becoming pregnant three months after method-related discontinuation. This indicates that, although they had not become pregnant, they had not switched to any other method of birth control. As switching is uncommon after three months, these women were at high risk of an unintended pregnancy. The highest rate was found in Kazakhstan (42.3%) and the lowest was in Peru (11.6%). This indicates that, across all countries, between 10-40% of women who discontinued their IUD due to dissatisfaction with the method were left without another method of family planning. a. Time to switching. The overall median estimate of switching to another modern method of contraception within three months of discontinuation was 49.9%. (See appendix Table A4 for country-specific rates). High levels of switching to another modern method were noted in Peru (70.5%) and Morocco (69.8%), with Bolivia and Kazakhstan both showing lower levels (16.7% and 25.2%, respectively), (Figure 5). The most popular destination method was the contraceptive pill in all countries except Indonesia and Peru where around 30% of women chose to switch to injectable contraception within three months of discontinuing with their IUD. Sterilisation was a popular choice for women in Latin America, namely Colombia (9.3%), Nicaragua (7.5%) and the Dominican Republic (5.7%). Switching back to an IUD (possibly a different type) was very rare except in Egypt and Kazakhstan. FIGURE 5: Switching behaviours three months post IUD discontinuation Percent 80 60 40 20 KK traditional DR BO EG modern TABLE 8: Status at three months after discontinuing IUD for method-related reasons. (Median values for the 14 countries) At risk Became pregnant Switched to: Modern methods All episodes 23.9 12.4 IUD Pill Injectable Barrier All modern reversible 0.5 24.6 8.8 8.2 49.9 Sterilisation Traditional 0.5 11.5 Place of residence Urban 27.1 13.9 0 26 6.5 5.4 48.4 0.7 11.5 Rural 24 10 0 23.9 7.4 8.8 50.3 0 12.3 Primary or less 24.7 13.6 0 22.9 8.5 2.8 42.7 0.1 7.5 Secondary+ 20.6 10.2 1.3 25.6 9.4 8.7 52.9 0 14.4 Spacers 25.8 14.8 0 23.5 6.9 6.9 48.3 0 13.7 Limiters 25.3 8.4 0.7 26.5 10.5 6.4 48.3 0.9 12.6 Poor 24.5 12.1 0 22.6 5.1 5.6 42.9 0 9 Rich 23.6 11.2 0 25 9.9 7.5 52.3 0 14.2 Women education Motivation Wealth index *Note row percentages sum to 100% at the country level (see Appendix Tables A.5-A.8) JO NC CO PH TR BD VN ID MA PE Source: Demographic and Health Surveys A preference for switching from an IUD to another modern method after three months rather than to a traditional method was found to occur in all countries except in Bolivia and Turkey, where the opposite behaviour was observed. Across the 14 countries, the median percentage of switching to traditional methods was 11.5%. Three months after IUD discontinuation, around a third of women who had switched method in Bolivia, Turkey and Viet Nam were relying more on less effective traditional methods to control their fertility. b. Differentials. On average, urban women were more likely to be pregnant already or at risk of conception compared with women in rural areas. Of urban women, 27.1% had discontinued the IUD yet not taken up any other method, with this figure standing at 24% amongst rural women (Table A5). The corresponding figures for pregnancy were 10% and 13.9% for rural and urban women, respectively. Results from rural Egypt, rural and urban Kazakhstan and urban Dominican Republic indicated that around 40% of those who discontinued the IUD remained at risk three months later. 25 26 Results Marie Stopes International Results Marie Stopes International FIGURE 6A: Switching to a modern FP method within three months of IUD discontinuation, by residence FIGURE 6B: Switching to a modern FP method within three months of IUD discontinuation, by education FIGURE 6C: Switching to a modern FP method within three months of IUD discontinuation, by motivation for use Percent Percent 60 60 60 60 60 40 40 40 40 40 20 20 20 20 20 rural urban Figure 6a shows very slightly higher levels of switching to another modern method in rural areas than in urban areas (50.3% and 48.4% respectively). In rural areas of Bolivia, switching to a modern method was particularly low (18.2%), due to higher rates of non-use, pregnancy and use of traditional methods. Rural areas of Morocco, Indonesia and Peru saw switching of over 70% and lower levels of non-use. In urban areas, high levels of switching were observed in Philippines, Peru and Morocco (around 60% or above). While there were higher occurrences of switching to traditional methods in urban Bolivia (54.9%), there was no reported switching to modern methods, although it must be noted that the number of episodes in this category was very small. Of the less educated women, 24.7% on average reported remaining at risk of pregnancy following discontinuation compared with 20.5% of their better educated counterparts (Table A6). Similarly, the proportion of women already pregnant was higher among the less well educated women than among the better educated (13.6% versus 10.2%). At the country level, almost half of the lesser educated women in Dominican Republic and 41.7% of more educated women in Kazakhstan were at risk of pregnancy. Around 45% of less educated women in Bolivia and a fifth in the Philippines fell pregnant in the three months following discontinuation, with around 17% of more educated women in Bolivia, Jordan and Kazakhstan also reporting pregnancy. Percent FIGURE 6D: Switching to a modern FP method within three months of IUD discontinuation, by wealth primary or less secondary or more The levels of switching to modern methods of contraception was 10% higher among women who had received secondary level education compared with women who had only received primary education or less (52.9% and 42.7% respectively), (Figure 6b). Also, more educated women reported switching to traditional methods than less educated women (14.4% and 7.5% respectively). Among women who had only received primary education or less, the highest levels of switching to a modern method were observed in Peru (77.1%) and Indonesia (72.2%). While in Peru this high level of switching was accompanied by a low level of women reportedly being at risk of pregnancy, Indonesia saw more women discontinuing but not switching to any other method. Higher levels of switching to a new IUD within three months were observed in more educated women. A quarter of both spacers and limiters reported being at risk of conception three months after discontinuing the IUD, with levels above 40% observed amongst spacers in Dominican Republic and Kazakhstan (Table A7). Pregnancies were more common amongst spacers (14.8%) than limiters (8.4%), with levels particularly high amongst Kazakhstani spacers (20.9%) and limiters, and Bolivian limiters (27.2%). Percent spacer limiter The probability of switching to a modern method was found to be similar for both switchers and limiters with the median for the 14 countries lying at 48.3% in both groups (Figure 6c). Limiters in Peru and Morocco were particularly likely to switch quickly to modern methods (74.4% and 74% respectively). Bolivian and Turkish women who were motivated to limit family size favoured less effective traditional methods above modern methods. In Bolivia, 10.8% of limiters switched to a modern method of contraception, yet switching levels to a traditional method were more than double, at 25.2%. In Turkey, 30.4% of limiters switched to a modern method, but 38.3% were relying on a less effective traditional method. High rates of sterilisation post-discontinuation were observed among limiters in three Latin American countries, namely Colombia (20.9%), Dominican Republic (18.1%) and Nicaragua (18.9%). These were also the only countries to report sterilisation rates among women who were reportedly only spacing their births. poorer richer On average, little difference was observed between richer and poorer women in the probability of becoming pregnant after IUD discontinuation or remaining at risk of pregnancy (Table A8). Overall, Figure 6d indicates higher levels of switching to a modern method among wealthier women compared with their poorer counterparts (50.4% and 42.9% respectively). The same pattern was observed for switching to traditional methods, with 14.2% of wealthier women turning to these methods compared with 9% of poorer women. Country-specific results are difficult to interpret because of the small number of episodes reported, particularly by poorer women. If attention is restricted to countries with 50 or more episodes in both groups, it may be noted that in Egypt and Nicaragua, richer women were more likely to switch to a modern method than poorer women but the reverse was true in Bangladesh and Jordan. In urban areas, high levels of switching were observed in Philippines, Peru and Morocco (around 60% or above). 27 28 Results Marie Stopes International Results Marie Stopes International 4.5 Comparison of IUD and other modern methods In this section, IUD discontinuation, reported failure and switching is compared to these indicators for the other main modern reversible methods, pills, injectables and condoms, using data from the 14 study countries. Figure 7 shows all cause discontinuation at 12 months for these methods. Similar results were obtained for method-related discontinuation (not shown). A huge difference is apparent between IUDs and the other three methods. Whereas a little over 10% of women stopped using their IUD in the first 12 months of use, the corresponding estimate for the other methods was over 40%. The height of the box for IUDs is also lower than those for other methods, indicating less inter-country variability for the IUD. An alternative way of representing discontinuation is to estimate the median length of episodes and the results are shown in Table 9. In all 14 countries, the length of IUD episodes vastly exceeds that of other modern methods. For these other reversible methods, the median length of use was typically about ten months, although longer in Indonesia (33 months) and Viet Nam (22 months). In comparison, median lengths of IUD use exceeded 50 months in half the 14 countries and were less than 30 months in only three of them: Bangladesh (20 months), Dominican Republic (25.5 months) and Nicaragua (27 months). The comparison is extended in Figure 8, which shows 12-month reported failure rates. As noted earlier, the failure rate for IUDs for all 14 countries was 1.3 per hundred episodes. The equivalent median value for injectables was not much higher, but the height of the box for this method indicates a much greater range of failure rates across countries. For pills and condoms, failure rates were much higher, close to eight per 100 episodes. Finally, it is of interest to ascertain whether methodswitching differs between discontinuers of IUDs and of other modern methods (Figure 9). Half of IUD discontinuers switched to another modern method within three months compared with a little over 40% of discontinuers of other modern methods. The heights of the boxes for IUD and other method users indicate greater inter-country variability in switching for the former than the latter group. Thus generalisations about differences in switching behaviour should be made with caution. TABLE 9: Duration of modern methods and IUD episodes FIGURE 7: All-cause discontinuation rates at 12 months for the four main modern, reversible methods Rate per 100 episodes 80 70 60 50 40 30 20 Method failure IUD Pill IUD Pill Injectables Condom 10 Injectables Condom All reasons FIGURE 8: Discontinuation rates at 12 months due to failure of the four main modern, reversible methods Rate per 100 episodes Country Modern methods (MM) no. of episodes median use (months) IUD no. of episodes median use (months) Bangladesh 4,802 11.8 369 20 Bolivia 996 6.1 495 50.3 Colombia 17,676 11.8 2,843 53.6 Dominican Republic 8,910 10.1 684 25.5 Egypt 4,811 18 5,008 35.1 Indonesia 18,191 33.2 682 60+ Jordan 3,334 14.4 2,624 37.3 Kazakhstan 990 8.2 1,125 59.5 Morocco 6,856 17.6 506 42.3 Nicaragua 4,713 11 1,291 27.4 Peru 2,940 10 267 36.7 Philippines 2,958 14.8 304 60+ Turkey 1,555 11.4 1,046 50.7 Viet Nam 923 21.7 1,241 60+ Total (median) (4,024) (12) (865) (37) Modern methods include pill, injections, vaginal methods, condom and Norplant 16 12 8 4 IUD Pill Injectables Method failure Condom 29 30 Discussion and conclusions Marie Stopes International 5. Discussion and conclusions 5.1 Data limitations The main limitations of this analysis stem from the narrow range of available data. The decision was made to use nationally representative data from DHS enquires that collected contraceptive calendars. The number of countries that could be included was further narrowed by the need to have a minimum number of IUD episodes for each country in order to justify detailed analysis. The net result of these requirements was that only 14 countries could be studied. Though their geographical spread was reasonable, global generalisations need to be made with caution. It is particularly regrettable that China, which accounts for about half of all IUD users, could not be included in the study because no DHS has yet been conducted in China. It is also regrettable that no data were available for any industrialised country. While DHS data are generally considered to be of high quality, they inevitably have limitations. The main limitation in respect of this report concerns the recording of reasons for stopping a method. Motivations for stopping may be complex and multi-faceted and it is difficult to capture such complexity in a large, structured interview survey. Moreover, the recording of only one reason for stopping is obviously not ideal. The most significant limitation is the lack of countryspecific information on the nature and quality of family planning services and on the types of IUD that are used. Huge variations between countries in IUD use dynamics are apparent. However, in the absence of contextual information on services, interpretation is impossible. The report can identify countries with, for instance, unusually poor IUD use continuation but cannot proffer reasons or propose remedies. FIGURE 9: Percentage of women who switched to any modern method within three months post discontinuation due to method-related reasons, by method % switched to any modern methods 70 60 50 40 30 20 IUD Other modern © Marie Stopes International The case for vigorous promotion of IUDs is particularly strong in sub-Saharan Africa, the one remaining region where contraceptive prevalence remains low but where large future increases are projected. 5.2 Main findings This analysis has served to confirm and elaborate results that were already established from earlier studies, such as the low reported failure rate of IUDs and low discontinuation compared with alternative methods. It has also generated useful new evidence, for instance, by examining IUD use dynamics among women of different characteristics and by analysing discontinuation in the second and third year following IUD adoption. The main results are summarised and discussed below: a. Even within the 14 study countries, which were selected for reasonably high overall contraceptive use and a minimum of 300 IUD use-episodes, the contribution of the IUD to contraceptive protection is immensely variable. It accounts for about two percent of users in some countries but more than half of users in others. It is most unlikely that any biological or cultural factor can account for this variability. Rather, it reflects the policy choices about which methods to promote and biases in family planning services. Discussion and conclusions Marie Stopes International In view of the advantages of the IUD over alternative reversible methods in terms of its low failure rate and high continuation, the virtual absence of IUDs from the overall method mix in so many countries is deeply regrettable. The case for vigorous promotion of IUDs is particularly strong in sub-Saharan Africa, the one remaining region where contraceptive prevalence remains low but where large future increases are projected. It is equally strong in a country such as Bangladesh, where women typically reach their desired family size in their mid-twenties and ideally require long-acting methods for the next two decades in preference to hormonal methods that currently dominate contraceptive provision. b. While the IUD, along with voluntary sterilisation and implants, is well suited to the needs of couples who have all the children they desire, an important finding from this analysis is that about half of all IUD use appears to be motivated by the wish to space or postpone childbearing. In only one of the 14 countries did the proportion of episodes attributable to women who still had not reached their desired size of family fall below 40%. 31 32 Discussion and conclusions Marie Stopes International Discussion and conclusions Marie Stopes International IUD use is compatible with breastfeeding and thus has a potentially valuable role in post-partum contraceptive protection. Women with infants often have unusually frequent contact with health services and thus, to the extent that family planning is integrated with maternal and child health services, abundant opportunities exist for IUD counselling and provision in the post-partum period. c. In terms of urban-rural residence and household wealth, IUD users differ little from users of other modern methods. In five of the 14 countries, IUD users were less well educated than other method users. However, in the majority of countries, this difference was negligible or reversed. Any suspicion that family planning staff target under-privileged and less educated women for IUD insertions because they cannot be ‘trusted’ to use hormonal methods receives no support from these results. d. The main reasons for stopping IUD use imply dissatisfaction with the method, but more intensive studies are needed to gain further insight into precise concerns. In view of the high proportion of women who are using the method for postponement of births, the percentage of women who say they stopped in order to have another child is surprisingly low. As noted above, the reliability of self-reported reasons is uncertain. One might have expected high levels of discontinuation in the early months following insertion, succeeded by a long period of low discontinuation. However, the evidence suggests that this is not the case. The probability of abandoning IUDs for methodrelated reasons is approximately constant over the three years following adoption, although this pattern varies between countries. e. Huge inter-country variations in the probability of stopping IUD use were observed. After 24 months, more than 40% had stopped use in three countries compared with 20% or less in another three countries. No obvious link to the overall ‘popularity’ of IUDs and discontinuation is apparent and thus the reasons for variation in the willingness or ability of women to persist with the method are unclear. f. In contrast to pronounced inter-country differences, discontinuation does not vary appreciably by women’s schooling, household wealth or urban-residence. It may be concluded that the method is equally suitable for all socio-economic and residential strata. © Marie Stopes International © Marie Stopes International g. Following IUD discontinuation for method-related reasons, half of all women on average switched to another reversible modern method within three months and an additional 11.5% switched to a traditional method. In most countries, the pill was the most common destination method and voluntary sterilisation was rare except in three Latin American countries. The proportion of women who switch ranges from 40% to 80% across the 14 countries, indicating very great variability in access to, or acceptability of, alternatives to the IUD. Better educated women and those from wealthier households were more likely to switch than their counterparts. Very few women readopted an IUD. IUD discontinuers were slightly more likely, on average, to switch to another modern method than those who discontinued another modern method. IUD users experience fewer side effects and worries about the possible effect of the method on their health than other users. h. The analysis confirms the huge advantages of the IUD over alternative (mainly hormonal) methods in terms of reported failure and discontinuation in all 14 countries. Its low failure rate implies a marked advantage over other methods in countries where abortion is illegal and often unsafe. Whereas more than 40% of users of the pill, injectable methods and condoms stopped use within 12 months, the equivalent figure for IUDs was only 13%. Similarly, while the median length of uninterrupted use of IUDs is typically 30 or more months, the median for other methods is typically ten months. Thus an IUD offers, on average, three times the length of contraceptive protection offered by other modern reversible methods. The reasons for this contrast do not appear to stem from differences in the characteristics of women who choose specific methods nor are they readily explained by variations in spacing versus limitation motives. It is possible that women who decide to use an IUD are more committed to avoiding pregnancy than those who choose another method. It is also likely that IUD users experience fewer side effects and worries about the possible effect of the method on their health than other users. However, the most significant reason may be the obvious one. A user of pills, injectables or condoms has to make a conscious effort to persist with use, whereas an IUD user has to make a conscious effort to stop. Long-term adherence to any drug regime that requires regular pill-taking or injections is known to be poor and, in this regard, contraception may be no different. 33 34 Appendix Marie Stopes International Appendix Marie Stopes International 6. Appendix TABLE A.1: Cumulative cause-specific discontinuation probabilities (per 100 episodes) At 12 months All reasons Failure Method-related Want a child No further need At 36 months All reasons Failure Method-related Want a child No further need Bangladesh 37.3 0.4 32.2 2.5 4.4 Bangladesh 66 1.2 53.6 12.4 14.4 Bolivia 11.9 1.7 8.1 1.1 1.4 Bolivia 40 1.9 26.9 9.6 7.2 Colombia 16.7 2.7 12.7 0.9 1 Colombia 38.8 6.9 28.4 5.3 2.8 Dominican Republic 28.2 2.5 19.7 2.7 5.5 Dominican Republic 63.1 4 46.5 18.4 11.4 Egypt 11.8 1 6.3 3.4 1.4 Egypt 51.2 3.5 18.4 34.6 4.6 Indonesia 10.4 0.7 6.8 0.7 1.7 Indonesia 24.6 4.9 10.9 6.8 3.2 Jordan 12 0.8 7.5 3.6 0.5 Jordan 48.2 5.5 20.3 28.7 3.2 Kazakhstan 11.4 3.2 6 1.5 1.2 Kazakhstan 30.5 6.3 15.4 9.9 2.7 Morocco 15.5 1.5 11.4 0.9 2.2 Morocco 43 1.9 27.1 16.9 3.8 Nicaragua 24.2 1.8 18 2.9 3 Nicaragua 58.1 6.1 40.9 11.7 13.7 Peru 20.1 0.7 17.5 2.1 0.4 Peru 47.1 3.8 34.4 11 5.7 Philippines 13.8 0.6 8.8 0.4 4.4 Philippines 33.3 0.6 27.2 2.7 5.2 Turkey 9.6 1 6.4 0.5 1.9 Turkey 30.8 3.3 16.1 7.3 7.8 Viet Nam 12.5 1.8 8.9 1.3 0.8 Viet Nam 30.6 4.9 16.4 11.1 1.7 Median 13.2 1.3 8.9 1.4 1.6 Median 41.5 3.9 27 11.1 4.9 At 24 months All reasons Failure Method-related Want a child No further need Bangladesh 53.9 1.2 46.5 6 6.8 Bolivia 25.7 1.9 15.9 5.5 4.6 Colombia 27.7 4.8 20 2.9 2.1 Dominican Republic 48.5 4 32.6 11.3 9.8 Egypt 33.5 2.4 12.7 19.3 2.9 Indonesia 15.3 1.9 8.3 2.2 2.6 Jordan 27.3 2.4 14 12.4 1 Kazakhstan 22.3 5.6 11.7 4.5 2.1 Morocco 29.7 1.9 20.2 7.7 2.6 Nicaragua 44 4.6 29.3 8.2 9 Peru 32.6 1.1 27.8 4.9 0.5 Philippines 23.2 0.6 17.5 1.1 5.2 Turkey 17.7 2.2 10.1 2.3 4.1 Viet Nam 20.4 3.6 12.6 4.3 1.2 Median 27.5 2.3 16.7 5.2 2.8 © Marie Stopes International 35 36 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.2: Conditional 12-month cause-specific discontinuation probabilities (per 100 episodes) Still using at 12 months All reasons Failure Method-related Want a child No other need Still using at 24 months All reasons Failure Method-related Want a child No other need Bangladesh 26.5 0.8 21 3.6 2.5 Bangladesh* - - - - - Bolivia 15.6 0.2 8.5 4.5 3.2 Bolivia 19.2 0 13.1 4.3 2.8 Colombia 13.2 2.2 8.4 2 1.1 Colombia 15.3 2.1 10.5 2.5 0.7 Dominican Republic 28.2 1.6 16 8.8 4.5 Dominican Republic 28.5 0 20.6 8 1.8 Egypt 24.6 1.4 6.9 16.4 1.6 Egypt 26.6 1.1 6.5 19 1.7 Indonesia 5.5 1.3 1.7 1.6 0.9 Indonesia 11 3 2.8 4.7 0.6 Jordan 17.4 1.6 7 9.2 0.5 Jordan 28.8 3.1 7.4 18.6 2.2 Kazakhstan 12.2 2.5 6.1 3.1 1 Kazakhstan 10.6 0.7 4.2 5.6 0.6 Morocco 16.9 0.4 10 6.9 0.3 Morocco 18.9 0 8.6 10 1.3 Nicaragua 26.1 2.9 13.8 5.4 6.2 Nicaragua 25.1 1.5 16.4 3.8 5.2 Peru 15.6 0.5 12.5 2.9 0.1 Peru* - - - - - Philippines 10.9 0 9.5 0.7 0.8 Philippines 13.2 0 11.8 1.6 0 Turkey 8.9 1.2 3.9 1.8 2.3 Turkey 16 1.1 6.8 5.1 3.9 Viet Nam 9.1 1.8 4 3.1 0.4 Viet Nam 12.8 1.4 4.4 7.1 0.5 Median 15.6 1.4 8.5 3.4 1.1 Median 17.5 1.1 8 5.4 1.5 *Rates of less than 100 episodes were suppressed TABLE A.3: Method-related discontinuation probabilities (per 100 episodes), disaggregated by selected characteristics At 12 months Residence Rural Urban Education Primary or less Secondary or more Motivation Spacer Limiter Wealth Poor Rich Bangladesh ­- 33 34.5 26.1 38.1 29 33.9 28.6 Bolivia 9.2 ­- 8.9 7.9 7.9 8.2 ­- 9.6 Colombia 11.8 15.8 10.5 13.4 15.6 9.6 17.7 10.8 Dominican Republic 18.4 24.3 26.3 16.9 20 19.3 ­- 14.2 Egypt 5.2 7.2 6 6.4 8.6 3.8 7.7 4.7 Indonesia 8.6 3 0.8 8.1 11 2.8 ­- 8.5 Jordan 7.2 8.9 6.2 7.5 8.2 6.5 8.1 6.9 Kazakhstan 5.2 6.9 ­- 6 6.9 4.3 7.4 5.5 Morocco 10.8 12.7 11.7 10.5 11.7 11.1 10.2 11.3 Nicaragua 18.2 17.1 17.6 18.2 19.9 15.4 18 17.8 Peru 17.6 ­- ­- 18 25.2 9.9 ­- 13.7 Philippines 4.8 11.7 ­- 9 13.1 5.6 9.2 ­- Turkey 5.9 7.9 6.6 5.9 8 5.3 7.5 5.1 Viet Nam 5.6 9.4 12 7.4 11.1 7.5 10.9 4.2 Median 8.6 10.6 10.5 8.6 11.4 7.9 9.7 9.6 Note: Rates of less than 100 episodes were suppressed © Marie Stopes International 37 38 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.4: Status at three months after discontinuing IUD for method-related reasons All episodes No. episodes* At risk Became pregnant Switched to: Modern methods IUD Pill Injectable Barrier All modern reversible Sterilisation Traditional method Total % Bangladesh 184 24.8 8.7 0.5 35.7 10.4 9 55.6 0.9 10 100 Bolivia 82 31 23.7 1.3 12 2 1.5 16.7 1.5 27.1 100 Colombia 599 21 7.9 0 23.4 17.6 10.7 51.7 9.3 10.2 100 Dominican Republic 183 39.7 12.6 0 26.1 7.2 3.6 36.8 5.7 5.1 100 Egypt 654 34.1 16.3 4.4 31 12.3 1.1 48.7 0 0.9 100 Indonesia 64 21.2 8.6 0 25.7 28.1 9.2 63.1 0 7.2 100 Jordan 432 28.3 17.2 1.7 23.6 0.8 9.2 35.1 0.1 19.3 100 Kazakhstan 178 42.3 17.3 6 8.9 2.2 8.2 25.2 0 15.1 100 Morocco 143 16.8 2.8 0 57.8 6.6 5.4 69.8 0.2 10.5 100 Nicaragua 366 26.5 8 0.5 37 13.9 2.9 54.3 7.5 3.7 100 Peru 68 11.6 5.1 0 21.4 30.5 18.5 70.5 0 12.8 100 Philippines 62 22.5 14.1 0 32.3 17.2 1.5 51 0 12.4 100 Turkey 155 23.1 12.2 1.9 19.9 1.8 8.1 31.7 0.9 32.1 100 Viet Nam 194 16.5 14.6 2.4 19 0.3 16.2 37.8 1.6 29.5 100 Median 181 23.9 12.4 0.5 24.6 8.8 8.2 49.9 0.5 11.5 *Discontinued due to method-related reasons © Marie Stopes International 39 40 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.5: Status at three months after discontinuing IUD for method-related reasons: by residence Rural No. episodes* At risk Became pregnant Switched to: Modern methods IUD Pill Injectable Barrier All modern reversible Sterilisation Traditional method Total % Bangladesh 29 17.9 9.1 0 28.1 8.8 14.4 51.3 0 21.8 100 Bolivia 75 31.4 25.8 1.5 13 2.1 1.6 18.2 0 24.6 100 Colombia 436 18.6 7.8 0 23.2 16.4 12 51.5 9.7 12.3 100 Dominican Republic 122 37.8 10.9 0 25.2 10.7 4.5 40.4 4 7 100 Egypt 259 37.2 11.3 8.5 30.7 7.4 2.7 49.3 0 2.2 100 Indonesia 42 15.8 3 0 22.4 37.9 10.3 70.6 0 10.6 100 Jordan 367 26.3 17.2 1.4 24.6 0.9 9.8 36.7 0 19.8 100 Kazakhstan 100 40.8 19.1 5.1 11.9 0.6 11.2 28.8 0 11.3 100 Morocco 103 14.6 2.2 0 59.2 8.8 5.5 73.5 0 9.7 100 Nicaragua 282 24.7 8.9 0.4 37.1 15.2 3.2 55.9 6 4.6 100 Peru 58 13.3 3.5 0 22.3 30.1 17.7 70.1 0 13.2 100 Philippines 21 35.3 18.1 0 30.1 4.2 0 34.3 0 12.3 100 Turkey 107 23.3 11.7 1.4 21.5 1 7.9 31.7 1.3 32.1 100 Viet Nam 17 21.2 7.1 0 20.6 0 32.6 53.2 0 18.5 100 Median 102 24 10 0 23.9 8.1 8.8 50.3 0 12.3 Urban No. episodes* At risk Became pregnant Switched to: Sterilisation Traditional method Total % Modern methods IUD Pill Injectable Barrier All modern reversible Bangladesh 154 26.1 8.7 0.6 37.1 10.7 8 56.4 1.1 7.7 100 Bolivia 7 26.9 0 0 0 0 0 0 18.2 54.9 100 Colombia 163 27.3 8 0 24.1 21 7.3 52.3 8 4.3 100 Dominican Republic 61 43.6 16.1 0 27.8 0 1.8 29.7 9.2 1.4 100 Egypt 395 32.1 19.6 1.7 31.1 15.5 0.1 48.4 0 0 100 Indonesia 22 31.9 19.3 0 32.3 9 7.1 48.4 0 0.4 100 Jordan 65 39.1 17.2 2.8 18 0 5.8 26.5 0.8 16.4 100 Kazakhstan 78 44.3 15.1 7.1 5 4.1 4.5 20.7 0 20 100 Morocco 41 22.3 4.2 0 54.3 1.1 5.1 60.5 0.5 12.5 100 Nicaragua 84 32.7 4.9 1 36.4 9.7 1.9 49 12.7 0.8 100 Peru 11 2.7 14.3 0 16.9 32.7 23 72.6 0 10.5 100 Philippines 41 16 12 0 33.5 23.8 2.3 59.6 0 12.5 100 Turkey 49 22.6 13.4 3.2 16.4 3.7 8.5 31.8 0 32.2 100 Viet Nam 177 16.1 15.3 2.6 18.8 0.3 14.6 36.3 1.8 30.5 100 Median 63 27.1 13.9 0.3 26 6.5 5.4 48.4 0.7 11.5 *Discontinued due to method-related reasons 41 42 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.6: Status at three months after discontinuing IUD for method-related reasons: by education Primary or less No. episodes* At risk Became pregnant Switched to: Modern methods IUD Pill Injectable Barrier All modern reversible Sterilisation Traditional method Total % Bangladesh 132 27.7 8.5 0 36.7 11 7 54.8 1.3 7.7 100 Bolivia 17 14.3 45.9 0 7.2 7.2 0 14.3 0 25.4 100 Colombia 156 22.6 13.5 0 20.9 15.6 6.3 42.8 12.4 8.7 100 Dominican Republic 64 48.4 14.4 0 24 0 1.1 25 8.8 3.4 100 Egypt 204 38.9 18.6 1.8 23 17.2 0.6 42.5 0 0 100 Indonesia 5 27.8 0 0 22.8 30.4 18.9 72.2 0 0 100 Jordan 20 34.4 13.6 0 19.8 5 0 24.8 0 27.3 100 Kazakhstan 2 100 0 0 0 0 0 0 0 0 100 Morocco 105 22.4 2.8 0 56.4 6.8 4.1 67.3 0.2 7.3 100 Nicaragua 159 24.2 11.3 0 40.4 9.9 1.5 51.7 10.9 1.8 100 Peru 12 1.8 11.9 0 4 73.1 0 77.1 0 9.3 100 Philippines 18 30.4 21.2 0 28.1 9.9 5.2 43.2 0 5.1 100 Turkey 124 25.2 15.1 1.3 19.9 2 5.5 28.7 1.1 29.9 100 Viet Nam 82 20.3 17.4 1 23.7 0.6 16.1 41.4 2.1 19 100 Median 73 26.5 13.6 0 22.9 8.5 2.8 42.7 0.1 7.5 Secondary or more No. episodes* At risk Became pregnant Switched to: Sterilisation Traditional method Total % Modern methods IUD Pill Injectable Barrier All modern reversible Bangladesh 52 17.3 9.3 1.8 33 8.8 14 57.6 0 15.8 100 Bolivia 65 35.4 17.9 1.7 13.2 0.6 1.9 17.4 1.9 27.5 100 Colombia 443 20.5 5.9 0 24.3 18.3 12.2 54.8 10.7 8.2 100 Dominican Republic 119 35.1 11.7 0 27.2 11 5 43.1 4.1 6 100 Egypt 450 31.9 15.2 5.6 34.6 10 1.4 51.6 0 1.3 100 Indonesia 60 20.7 9.2 0 26 27.9 8.5 62.4 0 7.7 100 Jordan 412 28 17.4 1.7 23.7 0.6 9.6 35.6 0.1 18.9 100 Kazakhstan 176 41.7 17.5 6 9 2.2 8.3 25.5 0 15.3 100 Morocco 39 1.6 2.6 0 61.6 6.1 9 76.7 0 19.2 100 Nicaragua 207 28.3 5.4 0.9 34.3 17.1 3.9 56.2 4.9 5.2 100 Peru 56 13.7 3.7 0 25.1 21.6 22.4 69.1 0 13.5 100 Philippines 43 19.2 11.1 0 34.1 20.2 0 54.3 0 15.4 100 Turkey 31 14.5 1.1 4.7 19.7 1.1 18.1 43.6 0 40.8 100 Viet Nam 113 13.8 12.6 3.4 15.6 0 16.2 35.2 1.4 37.1 100 Median 89 20.6 10.2 1.3 25.6 9.4 8.7 52.9 0 15.4 *Discontinued due to method-related reasons 43 44 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.7: Status at three months after discontinuing IUD for method-related reasons: by motivation for use Spacer No. episodes* At risk Became pregnant Switched to: Modern methods IUD Pill Injectable Barrier All modern reversible Sterilisation Traditional method Total % Bangladesh 65 24.8 17.2 0 37.1 4.7 9.9 51.7 0 6.3 100 Bolivia 41 28.1 20.2 0 15.9 4 3 22.8 0 29 100 Colombia 360 22 8.5 0 26.9 19.4 10.5 56.8 1.4 11.4 100 Dominican Republic 141 44.7 14.8 0 20.6 9.2 3.9 33.7 2.1 4.8 100 Egypt 409 34 18.8 2.3 30.7 12.1 1.5 46.6 0 0.7 100 Indonesia 49 19.3 11.3 0 17.2 42 8.6 67.9 0 1.5 100 Jordan 281 28.1 19.2 1.6 20.9 0.7 11.1 34.3 0 18.4 100 Kazakhstan 125 41 20.9 3.3 11.1 1.1 6.7 22.2 0 15.9 100 Morocco 74 13.7 4.5 0 57 4.3 4.6 65.9 0 15.9 100 Nicaragua 224 27 9.1 0.5 39.2 14.2 3 56.9 2.2 4.9 100 Peru 39 7.6 2.6 0 20.4 27.2 19.9 67.5 0 22.3 100 Philippines 29 18.5 20.4 0 40.9 9.1 0 50 0 11.2 100 Turkey 73 26.9 14.8 0 26.1 0 7.1 33.2 0 25.1 100 Viet Nam 86 19 11.2 1.3 18.8 0 18.1 38.2 0 31.6 100 Median 80.3 25.8 14.8 0 23.5 6.9 6.9 48.3 0 13.7 Limiter No. episodes* At risk Became pregnant Switched to: Sterilisation Traditional method Total % Modern methods IUD Pill Injectable Barrier All modern reversible Bangladesh 119 24.8 4.1 0.8 34.9 13.5 8.5 57.7 1.5 12 100 Bolivia 41 33.9 27.2 2.6 8.1 0 0 10.8 2.9 25.2 100 Colombia 243 19.1 6.8 0 17.9 16.4 10.8 45 20.9 8.1 100 Dominican Republic 42 22.9 5.3 0 44.8 0 2.7 47.5 18.1 6.2 100 Egypt 253 33.3 11.8 7.7 30.5 15.1 0.5 53.7 0 1.2 100 Indonesia 26 36.2 0 0 30.9 11.8 6.6 49.2 0 14.6 100 Jordan 151 28.6 13.4 1.7 28.4 0.9 5.6 36.6 0.4 21.1 100 Kazakhstan 53 45.4 8.8 12.4 3.6 4.7 11.9 32.5 0 13.2 100 Morocco 69 20.1 0.9 0 58.6 9.1 6.2 74 0.3 4.7 100 Nicaragua 142 25.8 6.3 0.6 33.4 13.6 2.7 50.2 15.9 1.9 100 Peru 29 17.1 8.5 0 22.8 35 16.6 74.4 0 0 100 Philippines 32 26.2 8.4 0 24.5 24.5 2.9 51.9 0 13.5 100 Turkey 82 19.7 9.9 3.7 14.4 3.5 9 30.4 1.6 38.3 100 Viet Nam 108 14.5 17.3 3.3 19.1 0.5 14.7 37.5 3 27.8 100 Median 75.7 25.3 8.4 0.7 26.5 10.5 6.4 48.3 0.9 12.6 *Discontinued due to method-related reasons 45 46 Appendix Marie Stopes International Appendix Marie Stopes International TABLE A.8: Status at three months after discontinuing IUD for method-related reasons: by wealth status Poorer No. episodes* At risk Became pregnant Switched to: Modern methods IUD Pill Injectable Barrier All modern reversible Sterilisation Traditional method Total % Bangladesh 51 19.7 7.2 0 38.1 16.1 7.9 62 3.4 7.7 100 Bolivia 2 25.5 0 0 0 0 0 0 0 74.5 100 Colombia 192 23.5 8.7 0 26.5 19.2 8 53.7 8.1 6.1 100 Dominican Republic 25 68.3 10.4 0 21.3 0 0 21.3 0 0 100 Egypt 179 37.9 19.6 1.5 21.6 18.8 0.6 42.6 0 0 100 Indonesia 3 23.6 0 0 18.8 20.6 33.7 73.1 0 3.4 100 Jordan 98 27.9 26.4 1.9 26.8 1.2 10.1 40 0 5.6 100 Kazakhstan 82 42.4 20.9 6.7 3.1 3.9 0.7 14.4 0 22.3 100 Morocco 29 14.7 5.9 0 63.9 1.5 3.7 69.1 0 10.3 100 Nicaragua 61 41.6 5.2 0 34.3 6.2 1.1 41.6 11.7 0 100 Peru 8 0.7 19.4 0 13.1 40.9 11.6 65.6 0 14.2 100 Philippines 30 20.4 13.8 0 35.6 15.9 3.1 54.6 0 11.2 100 Turkey 46 28.6 21.7 3.4 11.8 3.9 7.4 26.4 0 23.3 100 Viet Nam 103 18 17.9 4.6 23.5 0.5 14.7 43.3 2.2 18.6 100 Median 48.5 24.5 12.1 0 22.6 5.1 5.6 42.9 0 9 Richer No. episodes* At risk Became pregnant Switched to: Sterilisation Traditional method Total % Modern methods IUD Pill Injectable Barrier All modern reversible Bangladesh 73 22 7.4 1.3 31.9 6.6 13.3 53.1 0 17.5 100 Bolivia 64 32.2 20.5 1.7 15.3 2.5 1.9 21.4 1.9 24 100 Colombia 238 17.5 2.9 0 24.6 13.4 13.5 51.5 14.1 14 100 Dominican Republic 104 34.6 12.1 0 25.5 12.5 5.7 43.7 4.7 5 100 Egypt 239 34.7 13.5 7.6 33.7 5.9 2.3 49.4 0 2.4 100 Indonesia 51 16.1 10.8 0 29.3 28.8 6.7 64.9 0 8.2 100 Jordan 182 27.7 15.7 0.1 23.3 0.3 8.2 31.9 0 24.7 100 Kazakhstan 49 35.7 11.6 8.4 17.8 1.3 14.7 42.2 0 10.6 100 Morocco 67 18.1 1.5 0 53.9 7.2 8.3 69.4 0 11.1 100 Nicaragua 155 25.2 4.8 0.7 35.2 22.2 0.8 58.9 6.4 4.8 100 Peru 34 13 0 0 27.5 26.8 18.2 72.5 0 14.5 100 Philippines 14 29.3 8.1 0 16.8 21.8 0 38.6 0 24 100 Turkey 50 19.4 12.1 0 23.4 0 3.3 26.7 0 41.8 100 Viet Nam 25 16.3 15.1 0 18.3 0 35.5 53.8 0 14.8 100 Median 66 23.6 11.2 0 25 6.9 7.5 50.4 0 14.2 *Discontinued due to method-related reasons 47 48 Appendix Marie Stopes International 7. Glossary Definitions IUD episode: The duration in months of uninterrupted use of IUD that may or may not have ended by the time of the interview. Contraceptive calendar: This takes the form of a grid in which contraceptive status is recorded for each calendar month over a five year period preceding the survey. Time to first event: The time from start of use to discontinuation of the IUD. The single-decrement life-table: A hypothetical cause-specific probabilities in the absence of competing reasons for stopping also know as gross rates. Cu-IUD Copper-containing intrauterine device (IUD) DHS Demographic Health Survey FP Family planning hCG Human chorionic gonadotropin HIV Human immunodeficiency virus IUD Intrauterine device MEC Medical Eligibility Criteria for Contraceptive Use MM Modern methods PID Pelvic inflammatory disease STI Sexually transmitted infection TFR Total fertility rate WHO World Health Organization The multiple-decrement life table: The observed probabilities of cause-specific discontinuation in the presence of competing reasons, also known as net rates. Acronyms References 1.Richter R. Ein Mittel zur Verhütung der Konzeption [A means of preventing pregnancy]. Dtsch Med Wockenschr 1909;35:1525-7. 2.Kulier R, O’Brien PA, Helmerhorst FM, Usher-Patel M, D’Arcangues C. Copper containing, framed intra-uterine devices for contraception. Cochrane Database Syst Rev 2007(4):CD005347. 3.Trussell J. Contraceptive efficacy. In: Hatcher RA, Trussell J, Nelson AL, Cates W, Stewart FH, Kowal D. Contraceptive technology: nineteenth revised edition. . New York: Ardent Media, 2007. 4.United Nations. Levels and trends of contraceptive use as assessed in 2002. New York: United Nations, 2004. 5.United Nations. World Contraceptive Use 2009. New York: United Nations, 2009. 6.World Health Organization. Medical Eligibility Criteria for Contraceptive Use: Fourth Edition. Geneva: World Health Organization, 2009. 7.Goldman N, Moreno L, Westoff CF. Collection of survey data on contraception: an evaluation of an experiment in Peru. Stud Fam Plann 1989;20(3):147-57. 8.Westoff CF GN, Moreno L,. Dominican Republic experimental study; an evaluation of fertility and child health information. Calverton (MD): Macro International, 1990. 9.Strickler JA, Magnani RJ, McCann HG, Brown LF, Rice JC. The reliability of reporting of contraceptive behavior in DHS calendar data: evidence from Morocco. Stud Fam Plann 1997;28(1):44-53. 10.Goldman N WC. Can fertility be estimated from current pregnancy data? Population Studies 1980;34(3):535-50. © Marie Stopes International