In this study approximately 34% of women who smoked before pregnancy reported stopping smoking during pregnancy. This is slightly higher than figures reported in the 1999 – 2002/3 National Tobacco Strategy of 20 to 30%,  however since this time smoking cessation rates in pregnancy may have increased in line with the general community . In the analysis of the Australian Longitudinal Study on Women's Health year 2000 dataset a figure of 55% of women quitting was reported . Most recently however a figure of 4% of women quitting smoking during pregnancy has been reported by Moshin and Bauman  in a large cross sectional study in New South Wales, Australia. Internationally figures range from 15.8% of women quitting smoking during pregnancy from a national survey in Canada  to 26.8% from New Zealand . The disparity in these prevalence levels is most likely due to the timing of the data collection and whether the method of survey is cohort based or cross-sectional .
The relationship between smoking cessation during pregnancy and breastfeeding duration for longer than six months postpartum has not previously been reported in the research literature. More commonly continued maternal smoking in pregnancy has been reported in association with reduced breastfeeding initiation and duration [20-22], and only one previous study has explored smoking status and breastfeeding duration up to 26 weeks . In this study women who stopped smoking during pregnancy were significantly more likely to breastfeed for longer than six months, which is in accordance with national and international recommendations [24,25]. Although stopping smoking is not exclusively responsible for prolonged breastfeeding duration, [26,27] promoting smoking cessation during pregnancy supports both positive perinatal outcomes and supports optimal breastfeeding duration, known to be associated with protection against infection, some chronic diseases and improved cognitive development in the infant [25,28].
The reported effects of alcohol consumption as a predictor of smoking cessation during pregnancy are varied and inconsistent, however in this study consuming alcohol prior to pregnancy was significantly associated with stopping smoking during pregnancy. In previous research Severson et al  looked at alcohol in the week prior to the study questionnaire being administered (administered two weeks postpartum) and found that mothers who stopped smoking during pregnancy were less likely to have consumed alcohol in this week. Using data from the US National Maternal and Infant Health Survey, consuming one or more drinks during pregnancy was independently associated with a lower likelihood of quitting smoking during pregnancy  and in a study of the relationship between quitting tobacco, alcohol and caffeine consumption during pregnancy, Pirie et al found that quitting either alcohol or cigarettes was not associated with an increased likelihood of quitting the other substance (e.g. quitting alcohol was not associated with quitting smoking or vice versa). Although this relationship was not significant in the multivariate model the clustering of multiple substance use in individuals was . More recently a subset population of women in Spain who consumed alcohol (time of consumption not confined to either during pregnancy or three months after the birth) were also found to have a lower chance of quitting smoking .
In contrast, early research conducted in Sweden found continued alcohol consumption during pregnancy was not associated with a decrease in mothers stopping smoking . Similarly, data from Canada demonstrated that drinking during pregnancy was positively related to a woman's likelihood of attempting to quit smoking during pregnancy . However alcohol consumption was also significantly associated with cessation relapse before the child was born and the authors propose that although more women who drink make cessation attempts they are also more likely to relapse as it may be too difficult to give up smoking and drinking alcohol at the same time, or that continued alcohol use impairs the cessation maintenance. More recently, a New Zealand study found that women who quit smoking in the first trimester were more likely to report alcohol consumption at this time compared to women who reported not consuming alcohol .
In the current study alcohol intake before and during pregnancy was considered with regard to smoking cessation during pregnancy. The associations between alcohol before and during pregnancy with stopping smoking during pregnancy have previously not been studied concurrently. We found that women who consumed alcohol before pregnancy were more likely (OR = 2.6;95% CI 1.0–6.7; p [30,33] and therefore it is likely that those women who are consuming alcohol are also smoking hence the women most likely to stop smoking are those women drinking alcohol. Alcohol consumption during pregnancy was not significantly related to smoking cessation at this time.
In accordance with previous research, a woman was more likely to stop smoking during pregnancy if she was primigravida [16,18,29-31,34-38]. Women who have smoked during a previous pregnancy generally have an experience of giving birth to one or more healthy children and are therefore less motivated to quit smoking for subsequent pregnancies.
Pre-pregnancy smoking levels indicate that women who quit smoking during pregnancy are probably less addicted to smoking than women who continue to smoke. In the present study a woman was more likely to stop smoking if she reported smoking less than 10 cigarettes per day in the pre-pregnancy period. This result conforms with the current literature in that women who smoke at low levels are more likely to quit smoking [19,30,35-37,39,40].
Having a partner who smokes [19,31,35,36,38] and a low level of education  are factors previously found to be predictive of continued smoking during pregnancy. Although significant at the univariate level, education and father's smoking status were no longer significant when included in the multivariate analysis. Interestingly, a greater number of fathers stopped smoking after the baby was born. This may be due to the perception that the baby does not seem 'real' until after the birth when fathers are prompted by the baby's presence to quit smoking .
Antenatal classes aim to prepare expectant parents for childbirth and their new family life. Attendance at antenatal classes was a significant predictor of smoking cessation in the univariate analysis, although not significant in the multivariate model. Previous studies have found that early attendance at antenatal care was predictive of stopping smoking during pregnancy [16,35].
Timing of the pregnancy as a predictive factor for stopping smoking during pregnancy has not previously been reported using Australian data. Internationally previous research has shown that women having an unplanned pregnancy were more likely to continue smoking during pregnancy , whereas others have failed to find an effect of an unintended pregnancy . In this study, women whose pregnancy was unplanned were less likely to stop smoking during pregnancy, however this was not significant in the multivariate analysis. A planned pregnancy enables a woman the opportunity to consider stopping smoking in preparation for the antenatal period, whereas women who become pregnant unexpectantly have less time to implement this change.
Neither age nor income was related to the likelihood of stopping smoking during pregnancy. The correlation with age has been found in some previous studies  but not in others [38-40]. A relationship between age and smoking cessation during pregnancy is still unclear and further research is required in this area.
Studies have reported 66% higher medical costs attributed to complicated births for smoking mothers compared with non-smoking mothers . In Australia it has been estimated that smoking during pregnancy is responsible for 78 infant deaths, 6890 hospital separations and a cost of AUD23 million dollars to the health care system each year .
As smoking cessation programs have been shown to reduce the odds of continued smoking in pregnancy , it is imperative that the factors found in this study and previous research to predict smoking cessation during pregnancy be addressed in evidence based intervention programs. This concurs with recommendations from the 2001 National Tobacco Strategy . However despite this recommendation there appears to be a lack in the provision of any routine antenatal smoking cessation advice in the Australian health care setting [46,47].
This study did not define those women who quit in the pre-pregnancy period from those who quit during pregnancy, often referred to as 'spontaneous quitters' . In addition, there is considerable evidence outlining a high prevalence of relapse in the postpartum period in women who quit smoking during pregnancy,  however the design of this research study did not enable this issue to be addressed. Future cohort studies should take smoking abstinence into consideration in the design phase. Consideration of factors contributing to residual confounding, such as emotional antenatal attachment to the foetus in relation to smoking cessation, were not measured in this research. Future research should examine these additional potential factors that may help further explain the relationship between pregnancy and smoking cessation.
As in most studies of smoking during pregnancy, all smoking behaviours were self-reported in this study and cigarette smoking may have been underreported particularly during the antenatal period when there is an increased stigma associated with smoking. Nevertheless, self-reported smoking status is considered to be reasonably accurate [49,50] and results presented here give a good picture of smoking during pregnancy. Although this study used a standardised questionnaire, to elicit smoking information, future studies should consider the inclusion of alternative measures of cigarette smoking.
A further limitation of the study is having less than 60% of eligible women participate. Nevertheless, the sample size is still relatively large (>500), and there was no significant difference in maternal age and level of education between participant and non-participants, suggesting that the sample was representative of the population from which it was drawn.
Notwithstanding the relatively small sample size, and the fact that all women came from government-based hospitals, results from this study do reflect the current evidence. In addition these mothers are representative of the 'hard to reach' groups in Australian health promotion. Therefore the lessons learned from this study could be usefully applied in health education programs.