Socio-demographic determinants of intrauterine device use and failure in China

Duolao Wang,1 and Dan R. Altmann

Medical Statistics Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
BACKGROUND: This study examines social, demographic and family planning programme factors influencing intrauterine device (IUD) use, failure and subsequent resolution (`use dynamics') in the 1988 Chinese National Survey of Fertility and Contraceptive Prevalence. METHOD: A time-to-failure model was used to identify independent determinants of IUD failure. Logistic regression was used to identify independent predictors of abortion in women after failure. RESULTS: Being younger at IUD fitting [<25 versus 35 years, hazard ratio 5.9, 95% confidence interval (CI) 4.3, 7.7] and having a larger number of living children (>=3 versus <2 children, hazard ratio 1.2, 95% CI 1.1, 1.4) predict higher risk of IUD failure when controlled for each other; but in women with IUD failure, being older and having fewer children predict a much higher chance of resulting abortion. Contraceptive history and social/regional factors were also associated with higher IUD failure risk, in particular, use before 1984 (hazard ratio 1.3, 95% CI 1.2,1.4); and some of these factors were also predictive of abortion following failure. CONCLUSIONS: The determinants of IUD use dynamics suggest two main possible mechanisms. Some determinants may reflect effects of the Chinese family planning programme; some may indicate women's physiological and biological reactions to IUD. Health implications and relevant policy recommendations are discussed.

Key words: abortion/contraceptive failure/IUD/socio-demographic factors


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
Over 60 million women are using intrauterine devices (IUD) for contraception in China, where IUD is the most widely used form of contraceptive and also one of the main methods promoted by the national family planning programme (State Family Planning Commission, 1993; Wang, 1995Go). In 1988, IUD accounted for ~42% of contraceptive use (Wang, 1995Go). In 1992, contraceptive use prevalence rate reached an historic high of 83% among married women of reproductive age, with users of IUD still making up 40% of total users (State Family Planning Commission, 1993).

Although China has the largest number of IUD users in the world, there has been evidence that IUD failure rate in China is also among the highest. It has been shown that the first year failure rate for IUD is ~10% (Wang, 1995Go). This figure is the second highest after Brazil (13%) among developing countries, whereas in developed countries the first-year failure rate for IUD is only 2% (Moreno and Goldman, 1991Go; Hatcher et al., 1992Go). Some studies suggest that the high failure rates are related to the quality of IUD manufactured in China. Almost all IUD supplied in the 1980s were locally produced. The main IUD manufactured locally is the stainless steel ring (SSR) with several sizes (20–23 mm). There are also other locally designed devices available, including the `Mahua' ring (Copper V-200), and others (Xiao, 1989Go; Kaufman, 1993Go). One clinical study by the State Family Planning Commission of China in 1983–1984 in 13 provinces compared three IUD available in China—the SSR, the Copper T-220, and Copper V-200. It was found that SSR has a first year failure rate of 10.6%, compared with a rate of 1.7% for the Copper T-220 and 2.1% for the Copper V-200 (State Family Planning Commission, 1985). Another clinical study (Gao et al., 1986Go) in 1981–1986 in Beijing found a 2 year failure rate of 8% for the SSR, 2.1% for the Copper V-200 and 1% for the Copper T-220. However, a recent study (Tu, 1995Go), based on data from a specially-designed retrospective survey in four counties in North China, showed that the failure rates are higher for copper IUD but lower for non-copper IUD than those from the clinical studies and so the difference in the failure rates between copper and non-copper is unclear.

High failure rates of IUD used in China may have very important health implications. Induced abortion has been an important back-up method for the Chinese Family Planning Programme, and ~70% of the estimated 11 million abortions carried out each year were due to contraceptive, especially IUD, failure (Wang, 1995Go). Induced abortions, especially repeated induced abortions, may harm women's health due to operation-related morbidity and mortality.

Although the studies mentioned above provide us with some very important information about IUD failure in China, they have been largely limited to examining the effects of types or quality of IUD on their failure based on clinical studies, local studies or specialized surveys. The results from clinical studies are unlikely to reflect accurate information on the effectiveness of the method under real-life situations in China because of selection biases in subject recruitment, service quality, etc. Also, studies of IUD use and failure rates for the whole nation as well as for subgroups of the population are virtually non-existent.

In this paper we study the pattern and correlates of IUD use, failure and its subsequent resolution (sometimes referred to as use dynamics) in the 1980s under real-life situations using a national sample. The purposes of this study are 3-fold: (i) to describe the differentials in using IUD by some selected women's socio-demographic characteristics; (ii) to provide detailed information about the relationship between social, demographic, contraceptive use experience and family planning programme factors, and IUD failure; and (iii) to assess the impact of relevant variables on the decision to have an abortion if an IUD failure occurs.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
The data for this analysis come from a 10% subsample of the National Survey of Fertility and Contraceptive Prevalence, often referred to as the `Two per Thousand Fertility Survey', which was conducted by China's State Family Planning Commission from July 1, 1988 to July 15, 1988. The survey, representing a sample of two per 1000 persons in Mainland China, targeted ever-married resident women aged 15–57 years old. All provinces in the Chinese mainland took part in the survey.

A single-stage disproportionate stratified cluster sample design was used to select 13966 sampling points throughout the country. The sampling point was a small group with an average of 150 persons resident. The sample was stratified by province. The sample fraction was 1.98 per 1000 for the whole mainland population. Within provinces, however, various sample fractions were taken with the aim of obtaining representative data for each province. The average number of persons sampled per province was 72 000; altogether 2.15 million individuals were surveyed (Li, 1991Go; Lavely, 1991Go).

The central elements of the survey were pregnancy and contraceptive use histories, but data were also collected on the husband and other household members. In the contraceptive history, a woman was asked the month and year she had started practising contraception for each continuous period of contraceptive use, the method she had adopted, the date she had stopped using that method and the reasons she stopped at that specific time. A total of 27 contraceptive methods were listed in the questionnaire and this study focuses on IUD.

The sample for calculation of use prevalence is restricted to currently married women between the ages of 15 and 49 years at the time of survey (Jejeebhoy, 1990Go). This results in a sample of 40 226 women. The unit of analysis for IUD failure is the use-interval and a woman may contribute more than one use-interval. For this analysis, the sample is limited to all IUD use-intervals that began after 1 January 1980 to reduce recall errors and to obtain current information. The resulting number of IUD use-intervals is 14639, contributed by 12 334 women.

IUD failure is here defined as unintended pregnancy occurring while IUD is in use. This definition of failure includes both method failure and failures attributed to inconsistent or incorrect use, and is also called use failure (Jejeebhoy, 1990Go). A single decrement life table is used to estimate the probabilities of IUD failure by selected background characteristics. A proportional hazards model is used to analyse the simultaneous effects on the risk of IUD failure of a number of co-variates (Cox, 1972Go; Schirm et al., 1982Go; Grady et al., 1986Go). To control for the possible correlation among failures contributed by the same woman, a robust estimate of variance (Huber, 1967Go) is employed. Abortion rate, i.e. percentage of IUD failure ending in abortion, is also calculated for a number of selected variables and predictors of IUD failure resulting in an abortion against other outcomes of pregnancy (live birth, still birth or spontaneous abortion) are identified by a logistic regression model. For multivariate analysis, a combination of forward selection and backward elimination is utilized to obtain the most parsimonious model.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
IUD use prevalence
Table I Gopresents the contraceptive use prevalence rate by method and residence. Overall, 71.1% of married women aged 15–49 years were using a contraceptive method at the time of the survey in 1988. The most prevalent method in China was IUD with a prevalence rate of 29.6%; 41.6% of contraceptive users were fitted with IUD. IUD use is also characterized by urban–rural differentials. Urban women had a higher prevalence rate of 40.8%, compared with 26.8% for rural women. For women living in urban areas, the IUD was the most popular method, which accounts for more than half of total IUD use. Sterilization was the most prevalent method for the women in rural areas.


View this table:
[in this window]
[in a new window]
 
Table I. Percentage of currently married women aged 15–49 years using contraception by method and residence: 1988
 
IUD use by women's characteristics is shown in Table IIGo. The prevalence of IUD use increased with age up to ages 25–29 years, but declined thereafter, which may reflect a switch from IUD to other methods such as sterilization.


View this table:
[in this window]
[in a new window]
 
Table II. Intrauterine device (IUD) use prevalence rate by selected women's characteristics in 1988
 
Women with no living children had the lowest level of IUD use of 0.7%. This may be explained in part by a high proportion of women in this group being pregnant or wanting to get pregnant. IUD use was highest for women with one living child (almost 57%) but declined sharply to <26% for those with two, or more than two, children, who probably switched to sterilization (Wang, 1995Go).

There was little difference in IUD use between the majority Han nationality and the cultural minorities. In terms of regional differential in IUD use, the Northeast had the highest IUD use rate where the family planning programme is more effective.

IUD failure rate
Table III Godisplays the life table cumulative probabilities of contraceptive failure for 12, 24 and 36 months after initiating insertion of IUD, by a number of selected women's characteristics. Overall, IUD failure rates were 10.3, 16.3 and 20.2% within 12, 24 and 36 months respectively. This means that one in ten IUD users ended up with a pregnancy within the first year of IUD insertion and one in five during the first 3 years of IUD use.


View this table:
[in this window]
[in a new window]
 
Table III. Intrauterine device (IUD) failure rate by selected women's characteristics
 
The age of women at the start of IUD use had a marked effect on the likelihood of IUD failure. Failure declined as age increased, particularly after the 25–29 year age group. Table IIIGo also shows some differences in failure rates by number of living children, urban–rural residence, ethnicity, education, occupation and region of residence, but these differentials were not as pronounced as those by age.

Co-variate analysis of IUD failure
In the above analysis, simple life table techniques were used to describe the differentials in IUD failure rate by some background characteristics. The results from this tabulation, though interesting, do not separate the effects of each of these factors on IUD failure simultaneously. Another problem is that only the IUD failure during the first 3 years has been taken into account and little is known about failure after 3 years. These drawbacks can be overcome by using a proportional hazards time-to-failure model. The model is a multivariate approach that can be used to analyse the effects of several factors (or co-variates) simultaneously.

The aim of the hazards model analysis here is to determine how the hazard (or chance) of IUD failure depends on the explanatory variables covering socio-demographic characteristics and contraceptive use history. The percentage distribution of these variables is presented in the first column of Table IVGo. Two variables are used here to reflect the influences of the family planning programme: response to the programme campaign as a motivation for use, and period of use.


View this table:
[in this window]
[in a new window]
 
Table IV. Estimated effects of co-variates on IUD failure from proportional hazards model
 
The estimated results are reported in Table IVGo separately for both a crude univariate model and a multivariate model of best predictors of failure. Of the 11 explanatory variables examined and presented in Table IVGo, only seven variables emerged as direct and independent predictors, and these are indicated in the right-hand column of Table IVGo and outlined below.

Age at the start of IUD fitting was the most important and independent predictor of IUD failure. The results indicate that older women had a significantly lower IUD failure hazard. For example, the hazard for women aged >=35 years was only 17% of that for women aged <25 years.

After controlling for other variables, women with two or more children had ~20% higher failure hazard than women with no or one child. The crude model gives the impression of a reduced hazard for women with three or more children. The multivariate model shows that this is due to the greater age of these women: when women of similar age are compared, there is higher relative hazard due to three or more children, as shown in Table IVGo.

Women in rural areas had a 24% higher hazard than those in urban areas. Also, Table IVGo shows that observed regional differences were independent of the other explanatory factors; IUD failures were lowest in the Northwest region, which also had the lowest IUD use rate (Table IIGo).

Two variables representing contraceptive use history are shown to be associated with IUD failure. The crude model cannot disentangle the influence of previous contraceptive use from previous contraceptive failure, which of course implies use. As Table IVGo shows, controlling for all the other variables in the model gives final estimate of an 86% increased hazard with previous contraceptive failure, and an 18% hazard reduction with previous contraceptive use.

As Table IVGo shows, IUD fitted since 1984 are strongly associated with an IUD failure hazard 23% lower than those fitted before 1984.

Abortion as a resolution of IUD failure
Of various possible outcomes of IUD failure (such as live birth, stillbirth, spontaneous abortion), we consider here failures ending in abortion. To examine this outcome, the proportion of failures ending in abortion (abortion rate given IUD failure) is tabulated by some selected explanatory variables in Table VGo. The results of univariate and multivariate logistic regression models are also shown in Table VGo.


View this table:
[in this window]
[in a new window]
 
Table V. Proportion of intrauterine device (IUD) failures ending in abortion by selected women's characteristics and logistic regression analysis results
 
Among the 14 639 spells of IUD use, 2357 eventually ended in IUD failure within the study period. Failures occurring to those who were still pregnant at the time of the interview are excluded from this analysis. The sample thus consists of 2289 failures. Overall, 62.6% of IUD failures led to abortion.

The number of living children was the most important factor predicting whether an IUD failure was followed by abortion. Over half of IUD failures occurred to women who have one or no child and almost all these unintended pregnancies were aborted: abortion rate is 96.3% and the adjusted odds of abortion is 115.3-fold that of women with three or more children. The less pronounced univariate odds ratio in women with one or no child is due to not allowing for the lower odds in younger women: the abortion rate increased with the age of the woman.

The differentials in abortion rates were also notable for urban–rural residence and education. Urban women or educated women were more likely to have their unwanted pregnancies aborted. Women who had experienced a prior contraceptive failure also had a higher abortion rate than those without prior failure.


    Discussion and conclusion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
This study suggests that socio-demographic factors which in the West may not be expected to be associated with IUD efficacy can assume importance in a setting such as China.

The main feature of IUD use in China is its close association with family size. Age at start of IUD use, the number of living children, the period at start of IUD insertion, previous experience of contraceptive use and failure, region and urban–rural residence were all identified as significant independent determinants of IUD failure. The number of living children was again a very strong predictor of abortion given an IUD failure, as was age and to a lesser extent prior contraceptive failure, education and urban–rural residence.

The results from this study suggest two main possible mechanisms through which the co-variates included in the analysis may influence the dynamics of IUD use: some factors may reflect the effect of the family planning programme; other factors may illustrate women's physiological and biological reactions to IUD.

The first type of determinant may mirror the effect of the family planning programme on IUD use dynamics. The most obvious example is IUD choice and use. Chinese women of reproductive age seldom used IUD before their first child, but nearly half of the women who had just one child adopted IUD. When women had two or more children, only a quarter of them were still fitted with IUD whereas over half of them turned to sterilization (Wang, 1995Go). The choice of IUD use clearly indicates the implementation of the family planning programme since the policy strongly encouraged the insertion of an IUD after the birth of the first child, and sterilization after two or more children.

The effect of period of IUD use on IUD failure is a second example of the family planning programme effect. The results from this study show that IUD fitted since 1984 had a much lower failure rate than those fitted before 1984. The reason for this decline in IUD failure rate may perhaps be the promotion of the family planning programme. Another possible explanation may be relaxation of the family planning programme after 1983 as documented (Greenhalgh, 1986Go). This policy of leniency, when it happens, is likely to affect contraceptive behaviour. For example, users may be more motivated to use an IUD, so that the failure rate is lower. Furthermore, the Chinese government had been working very hard to promote the family planning services and more reliable types of IUD were introduced in the 1980s (Kaufman, 1993Go). Lower failure rate in the Northwest region and in urban areas observed in this study may also be related to the policy and programme. The Northwest region, mainly populated by ethnic minority, has a relaxed family planning programme, and the family planning services in urban areas are much better than rural areas (Wang, 1995Go).

Resolution of IUD failure by women with one child provides further evidence of the strong family planning programme impact. Once a woman with one child had an IUD failure, she would almost definitely have an induced abortion. This result indicates a strong commitment to the one-child policy by some Chinese women. An alternative interpretation could be that the programme puts pressure on these women to terminate their pregnancy.

Some other reasons of IUD failure and its subsequent resolution may also be essentially related to the family planning programme. Higher failure rate among women with two or more children is worth commenting on. In China, some women, especially multiparous mothers in rural areas, may have their IUD removed privately, but reported the pregnancy as a method failure, and then had a birth. In doing so, they could avoid incurring a penalty from the government (Wang, 1995Go). This point is further supported by the fact that some women, who already had two or more children, preferred to have a live birth rather than an abortion if they experienced an IUD failure. Those results may indicate the negative impact of the Chinese family planning programme.

Second types of correlates may reflect the action of physiological effects on IUD failure. IUD failure occurs more frequently among younger women. This is expected because a woman's exposure to the risk of pregnancy decreases with declining fecundity (Grady et al., 1986Go). Also, there may be lower coital frequencies at older ages (Grady et al., 1986Go). Another explanation may be that older women may be more experienced and motivated in the use of IUD and hence more efficient in their use of contraception to restrict their childbearing. Another example of physiological effect may be the effect of previous failure on the hazard of IUD failure. The hazards model results indicated that women with a prior contraceptive failure had almost twice the failure hazards than those without. This may suggest that IUD failure may occur repeatedly in some groups of women: a clustering of contraceptive failure. Whether this is due to women's physiological reaction to contraceptives needs further investigation.

We have suggested two possible mechanisms through which the variables selected in this study may affect the IUD use dynamics in China. Other mechanisms may also be in operation given that many other factors, observable and unobservable, which may affect the use dynamics, were not included in the analysis.

The results of this study have some reproductive health implications. According to the estimates from the 1993 survey data, there were >60 million IUD users in China. Therefore ~6 million women experienced a contraceptive failure during their first year of IUD use. Of the 6 million IUD failures, ~4 million would be aborted. The huge number of abortions caused by IUD failure only, mainly among young women, may increase the risk of disease and even operation-related death. Therefore, great efforts should be made to reduce IUD use failure for health reasons. The Chinese family planning programme should pay special attention to those IUD users who are young; who have no more than one child; and also those who have previous experience of unwanted pregnancy. Motivating users in their commitment to method use is also important to reduce the occurrence of IUD failure. The replacement of the cheap and widely used stainless steel ring with new IUD such as Copper-T may reduce levels of IUD failure. High quality and wide availability of IUD of various sizes and types should be provided for users. Training of family planning workers and the provision of improved information and counselling to clients are also essential.

This study has the limitation that the data used were collected over decade ago, and comprise the only resource available for the study of IUD use dynamics under real-life situations at the national level. As the Chinese family planning policy has hardly changed since the later 1970s, the results from this study could still be of importance to the Chinese family planning programme as well as to the understanding of the socio-demographic determinants of IUD use and failure in China.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
Thanks are due to the State Family Planning Commission of China for access to the 1988 National Survey of Fertility and Contraceptive Prevalence and to the Programme of Research Development and Training in Human Reproduction, World Health Organization, and the Population Council in New York, which partially funded this work.


    Notes
 
1 To whom correspondence should be addressed. E-mail: duolao.wang{at}lshtm.ac.uk Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion and conclusion
 Acknowledgements
 References
 
Cox, D.R. (1972) Regression models and life tables. J. R. Statist. Soc., Series B, 34, 187–220.

Gao, J., Shen, H., Zheng, S., Fan, H., Hui, M.H., Han, L.H. and Yao, G.Z. (1986) A randomised contraceptive clinical evaluation of steel ring, V Cu-200 and T Cu-220C IUDs. Contraception, 33, 443–454.[ISI][Medline]

Grady, W.R., Hayward, M.D. and Yagi, J. (1986) Contraceptive failure in the United States: estimates from the 1982 National Survey Family Growth. Fam. Plann. Perspect., 18, 200–209.[ISI][Medline]

Greenhalgh, S. (1986) Shifts in China's population policy, 1984–1986: views from the central, provincial, and local levels. Pop. Dev. Rev., 12, 491–515.[ISI]

Hatcher, R.A., Stewart, F., Trussell, J., Kowal, D., Guest, F., Stewart, G.K. and Cates, W. (1992) Intrauterine devices (IUDs). In Contraceptive Technology. Irvington, New York, pp. 355–385.

Huber, P.J. (1967) The behavior of maximum likelihood estimates under non-standard conditions. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Berkeley, CA. University of California Press, vol. 1, pp. 221–233.

Jejeebhoy, S. (1990) Measuring contraceptive use-failure and continuation: an overview of new approaches. In Measuring the Dynamics of Contraceptive Use: Proceedings of the United Nations Expert Group Meeting. United Nations, New York, pp. 21–45.

Kaufman, J. (1993) The cost of IUD failure in China. Stud. Fam. Plann., 24, 194–196.[ISI][Medline]

Lavely, W.R. (1991) China unveils its monumental Two per Thousand-Fertility Survey. Asian Pop. Forum, 5, 89–92.

Li, H.G. (1991) Introduction to national survey of fertility and contraceptive prevalence. In Proceedings of National Seminar on National Survey of Fertility and Contraceptive Prevalence. Hangzhou, China, pp. 2–9.

Moreno, L. and Goldman, N. (1991), Contraceptive failure rates in developing countries: evidence from the demographic and health survey. Fam. Plann. Perspect., 17, 44–49.

Schirm, A.L., Trussell, J. Menken, J. and Grady, W.R. (1982) Contraceptive failure in the United States: The impact of social, economic and demographic factors. Fam. Plann. Perspect., 14, 68–75.[ISI][Medline]

State Family Planning Commission of China (1985) A clinical trial of three commonly used IUDs. In Condensed Summary of Family Planning Scientific Research Results during the Sixth Five-Year Plan. State Family Planning Commission of China, Beijing, pp. 19–21.

State Family Planning Commission of China (1993) Data on Population and Family Planning. State Family Planning Commission of China, Beijing.

Tu, P. (1995) IUD discontinuation patterns and correlates in four counties in North China. Stud. Fam. Plann., 26, 169–179.[ISI][Medline]

Wang D. (1995) Contraceptive use dynamics in China. PhD thesis, Department of Social Statistics, University of Southampton.

Xiao, B. (1989) Contraception in China. Outlook, 71, 2–6.

Submitted on October 26, 2001; accepted on January 18, 2001.





This Article
Abstract
FREE Full Text (PDF )
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Request Permissions
Google Scholar
Articles by Wang, D.
Articles by Altmann, D. R.
PubMed
PubMed Citation
Articles by Wang, D.
Articles by Altmann, D. R.