|
nif:isString
|
-
Data come from six national probability surveys that explored lifetime contraceptive use and pregnancy intentions in France from 1978 through 2010 (Table 1) and from annual national statistics on abortions. All surveys studied national probability samples of several thousand women (range: 1,431 to 5,272) aged 18 to 44 years (the lowest age in the 1978 and 1994 surveys was 20). Sampling design and post-stratification weights were applied, using the same methodology in all surveys. Data are from national surveys and demographic statistics, so available upon request.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0091539.t001 Methodological characteristics of the 6 national surveys on contraceptive and reproductive behaviors in France. *Women were asked if they had planned to be pregnant in the 2000 survey whereas they were asked if they wanted to be pregnant in the other surveys. **The survey was offered to a national random sample who answered a mandatory governmental survey.
We present the distribution of contraceptive practices in 1978, 1988, 1994, 2000, and 2010 (Table 2). In case of dual use, we retained the most effective method, based on typical use failure rates published in the literature [12]. As 18- and 19-year-olds were not interviewed in 1978 and 1994, we used retrospective information collected in women’s lifetime contraceptive histories in the 1988 survey (among those aged 28 and 29 years) and in the 2000 survey (among those aged 24 and 25 years) to reconstruct their use of contraception in 1978 and 1994.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0091539.t002 Proportion of women 18–44 years old using selected contraceptive methods, by year. *Pill, IUD, Implant, Patch, Injections, vaginal ring.All percentages are weighted to account for sampling design and post-stratification.
To calculate annual contraceptive rates for women aged 18–44, we applied age-specific rates for each survey year to the age distribution of the female population in France during those years.
In each survey, women completed a lifetime pregnancy history. We classified each pregnancy as “wanted” if women stated they had wished to become pregnant “at that time” or “earlier”, and “unwanted” if they had wished to become pregnant “later” or “not at all” or if they stated they had “not thought about it”. Women who did not answer this question, i.e. between 0.5% and 3.2% of the women in the selected surveys, were excluded from the analysis. As abortions are underreported in population-based surveys [8], [13], we cannot use survey data to estimate unwanted pregnancies ending in abortion. Therefore, we estimated only unwanted births from the population-based surveys. We applied the proportion of wanted and unwanted births estimated from each survey to the national fertility rates (derived from national birth statistics) of the corresponding period to estimate national wanted and unwanted birth rates. National abortion rates (based on mandatory reporting of abortions in France) were then added to calculate the overall unwanted pregnancy rate as the sum of national unwanted birth and abortion rates (assuming all abortions were unwanted pregnancies). The probability of terminating an unwanted pregnancy was defined as the abortion rate divided by the unwanted pregnancy rate. To have a sufficient number of events to produce robust indicators, we estimated the proportion of births that were unwanted at the time of each survey as the average proportion of births that were unwanted over the 5-years period preceding each survey. We thus present unwanted pregnancy rates for the periods 1973–77, 1978–82, 1983–87, 1988–92, 1993–97, 1998–2002 and 2005-2009 (Table 3). Abortion rates represent the annual abortion rate of the median year of each interval.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0091539.t003 Estimates of lifetime pregnancy rates by year, age group, and pregnancy intentions. Results indicate that in 1973–77 women would have had 2.52 lifetime pregnancies according to the fertility age rates observed at that period. Among these pregnancies, 1.29 would have been wanted births, 1.23 unwanted pregnancies (0.59 unwanted births and 0.64 abortions). * Miscarriages are not included. (1) (2) Surveys data. (3) National statistcs. (4)National statistics (Pison 2010). (5) Surveys data (Beaujouan and Toulemon. 2013) for women aged 25–34.
Finally, we also present data on age at first birth based on national statistics [14] and on ideal family size based on national survey data [15]. We present the results for all women and for two age groups: 18–29 years and 30–44 years. To ensure that the social and demographic composition of the French female population for each survey year was accurately portrayed, sampling and post-stratification weights were used in each survey data set to take into account the complex survey designs and sampling distortions [16].
Estimation of rates over time: To estimate rates of contraceptive use and unwanted pregnancy for the same years, we used linear regression to model contraceptive use, known at the time of the surveys (from 1978 to 2010), and unwanted pregnancy rates, known for the median year of the period for each of the 5-years periods (from 1973–77 to 2005–09). We thus obtained predictive values of effective contraceptive prevalence rates and unwanted pregnancy rates for the following ten years: 1975, 1978, 1980, 1985, 1988, 1990, 1994, 1995, 2000, 2007 (Table 3). Abortion rates were available from national statistics for each year.
To test if the association between trends in effective contraceptive prevalence rates and unwanted pregnancy rates was stronger than the association between effective contraceptive prevalence rates and abortion rates, we explored the association between trends in effective contraceptive prevalence rates and trends in (i) abortion rates, (ii) unwanted pregnancy rates, and (iii) unwanted birth rates and used a multivariate linear regression [17] with three dependent variables (abortion, unwanted pregnancy and unwanted birth) to determine which of these three associations was strongest (Table 4).
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0091539.t004 Associations between contraceptives prevalence rates and abortion, unwanted pregnancy, and unwanted birth rates from 1973–77 to 2005–09. Results from multivariate linear regressions.
All analyses were performed with Stata statistical software (version 12.1).
All national surveys were approved for ethical and confidentiality issues by the French national data protection agency, the “Commission Nationale Informatiques et Libertés”.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}