The socio-demographic characteristics of the participants are shown on Table 1.

A relatively large proportion of the sample (214 out of 1660, or 12.9%) did not respond the question about history of having an unwanted pregnancy. One fourth (20.5%) of the respondents reported having had the experience of an absolutely unwanted pregnancy in themselves, if female, or with their partner, if male. Over half of those pregnancies were aborted, as shown in Table 2.

Almost one third of the respondents gave moral or religious reasons for not aborting an absolutely unwanted pregnancy and one fourth reported lack of knowledge, money or courage; these and the other reasons are shown on Table 3.

The most frequent reason for deciding to terminate the unwanted pregnancy was being single with 37%. Arround 30% aborted because the male partner did not accept that he was the father, for fear of their parents reaction and in order to keep studying. Almost one fourth felt that they were too young to become a mother or father. Only 6% referred to financial problems (Table 4).

The decision to abort the unwanted pregnancy was associated to not having a permanent partner, higher educational level and using barrier or behavioral methods of contraception at the time of getting pregnant, in bivariate analysis (Table 5).

The main reason for not using a contraceptive method at the time the woman got pregnant was that she did not expect to have a sexual relation (39% of cases) and in an additional 11% the male promised to use a condom, but did not used it. Only 5% did not have information on contraception. Respondents gave a variety of other reasons related to lack of responsibility and disbelief that they could get pregnant (Data not shown in tables).

The Poisson regression showed that the only factor being independently associated to terminate the unwanted pregnancy was having higher education (Table 6).

According to the answer of the interviewees, only 20% had experienced an absolutely unwanted pregnancy and only 11% had aborted. These percentages are much lower than the proportion of women who declared to have aborted found in other studies with similar populations, but where face to face interviews were used to collect the information11.

As abortion is basically illegal in Brazil, it is to be expected that an unknown proportion of those who responded a self reported questionnaire did not trust the confidentiality of the data collection and were not willing to declare to have done something that is punished as a crime in this country. The difficulty to obtain correct information on abortion has already been shown in our population 6.

Looking at the reasons given to terminate their pregnancy by those who declared to have aborted, it is clear that the reasons concentrate around answers which indicated that the pregnancy interfered with their life expectancy: to continue studying or working being single , too young to get married and to become a father or mother . Also important were indicators of family or partnership disfunction, such as fear of parents and denial of responsibility for the pregnancy by the partner. As reason to abort, the little relevance of finantial problems is a disagreement with a similar study carried out in 1990, where financial problems was given as a reason for abortion by almost three quarters of the respondents of a mail survey7. The low relevance of finantial problems in this sample may be explained by the relatively high socioeconomic status of the study sample, as judged by the high level of schooling, with more than two third of the sample population having higher education.

Among those who did not terminate an unwanted pregnancy it appears that religion acted as a barrier among almost one third of them and the legal barrier for almost one fourth. The effect of religion over opinion and behavior related to abortion has been already described in a number of studies on abortion and it is not a surprise 71213. The proportion of over 30% who did not abort for moral or religious reason is in agreement with the almost 70% of gynecologists who gave great relevance to religion and who declared to have aborted their own unwanted pregnancies14.

The fear of committing a crime punished by law is also understandable, and it may have also been a reason for the relatively low proportion of subjects who declared to have had an abortion. The high proportion of respondent, over three quarters, who did not gave this reason for not having a abortion is in agreement with the international acknowledgement that women who decide to terminate a pregnancy do so in a similar proportion independently of the legal conditions of abortion in the country they live15.

The significant association of higher educational level with a higher chance of having an abortion when confronted with an unwanted pregnancy is also in agreement with earlier studies, which shows that although women with higher education had less risk of having an unintended and unawanted pregnancy, when they get pregnant against their will, they do have a higher risk of aborting16.

The significantly higher proportion of abortion among those who were using a barrier or behavioral method of contraception, in comparison with those using none or modern methods was not expected, considering that behavioral methods are promoted by religious groups, which also oppose to abortion. However, as the correlation disappears in the multiple regression analysis it may be that it only reflects that women with higher education use behavioral and barrier methods more often than those with lower eduation17,18.

The main limitation of this study is that it is based in the data obtained from a self-responded questionnaire and it is impossible to estimate how well those who responded represent the universe of public servant we intended to reach. It means that the percentages of participants with history of unwanted pregnancy and abortion are not necessarily representative of the population of civil servant in the interior of the state of Sao Paulo that we intended to sample.

It doesn t mean, however, that the reason given for having or not having abortion is not valid, as that is not dependent on which proportion responded and on possible selection bias.

We also did not explore in this study, what happened with the babies born from unwanted pregnancies which were not ternminated. A study had suggested that those babies are at higher risk or morbidity and mortality that babies born from planned pregnancies19.

Our results showing that more than half of the respondents who have an unwanted pregnancy decided to terminate their pregnancies, contributes to confirm that the most important means to prevent abortion is by preventing unwanted pregnancies. Such purpose is achieved by facilitating access to information and provision of modern contraceptives, incluing long acting method given at no cost20, and also, by providing early education on responsible sexuality, which is known to delay sexual experience and to reduce the number of partners, unwanted pregnancies and abortions among adolescents21,22.

We hope that the publication of these results contribute to estimulate authorities to promote and carry out the interventions known to be effective in reducing unwanted pregananciess and abortion, instead of insisting in keeping abortion illegal, which has been shown to be ineffective to achieve that purpose15.