"Smoking ban 'cuts premature births'," BBC News has reported. Despite the BBC News headline, this research only showed an association between the smoking ban and a reduction in premature births. It didn’t show direct cause and effect.
The research that the BBC News story is based on recorded premature birth trends in Belgium around the time of a public smoking ban there. While not directly comparable to the bans in the UK countries, Belgium is a useful example to look at as they introduced public smoking bans in stages between 2006 and 2010.
The fact that there were three distinct steps means the beneficial impact of public smoking bans can be assessed more precisely.
The study found the number of preterm births dropped after each consecutive smoking ban, but can’t prove the smoking ban itself cut these rates. Other factors may also have been involved. For example, improvements in antenatal care may have reduced the rate of premature birth.
The fairest summary of the findings is that they provide some circumstantial evidence that smoking bans may reduce premature birth rates. They do not provide conclusive evidence of a link.
The study was carried out by researchers from the University of Leuven and Hasselt University in Belgium, and was funded by the Flemish Scientific Fund and Hasselt University.
The study was published as an open-access article in the peer-reviewed British Medical Journal.
Despite the slightly simplistic headline, the BBC News story offers an appropriate interpretation of the results. The BBC explains that the study found an association but could not prove that the ban was the cause of the observed drop.
Before and after studies are a simple way to assess the impact of policies. However, the fact that other factors or trends may have occurred at the same time as the policy was implemented can sometimes bias results.
This was an observational study (before and after study) looking into the association between a public smoking ban and the number of preterm births in Belgium.
Belgium’s smoking ban was introduced in three phases:
Smoking during pregnancy has been found to impair the baby’s growth and to be associated with preterm birth. Evidence relating to the effects of secondhand smoke exposure and risk of preterm birth is less consistent.
The researchers were interested to see if a smoking ban applied in phases across the region would be associated with the number of preterm births. Observational studies over several time points can be helpful in defining trends and links between two factors. If effects are large and interpreted alongside other studies they can build a case that one factor (in this case, public smoking bans) may be strongly linked to an outcome (preterm birth).
Researchers collected data on births in Flanders (a region in Belgium) from 2002 to 2011.
Births before 24 weeks’ gestation, after 44 weeks’ gestation and multiples births were not included in the analysis.
They researchers used these data to determine the annual risk of preterm birth in the years preceding the public smoking ban, during the three phases of the ban and immediately after the ban. They analysed the trend in this risk over time.
A second analysis was conducted to determine the percentage change in risk of preterm birth after the introduction of each phase of the smoking ban. Several potentially confounding factors were considered during this analysis, including:
Between 2002 and 2011, there were 606,877 births that were included in the study. Of these, 32,123 (7.2%) were classified as preterm births (occurring before 37 weeks’ gestation).
When examining the unadjusted percentage of births that were considered preterm, the researchers found that the rate in the four years prior to the smoking ban was relatively stable (although there was a slight reduction seen between 2004 and 2005).
After the first phase of the ban (2006 to 2007), the percentage of births classified as preterm dropped, and a further drop was seen in the year after the second phase ban (2007 to 2008).
A slight upturn was seen in early 2008, followed by another decline through 2009. After the third phase of the smoking ban was introduced in January 2010, an additional drop in the percentage of preterm births was seen.
When analysing the data while adjusting for the potential confounding factors, the researchers found that the risk of preterm delivery was reduced after each of the smoking ban introductions, with the drop being largest after the second and third phase of the bans.
After the second phase was introduced (banning smoking in restaurants), there was a 3.13% drop in the annual rate of spontaneous preterm delivery (95% confidence interval (CI) -4.37 to -1.87%). Following the third phase (no smoking in bars serving food) this drop in rate was -2.65% each year after January 2010 (95% CI -5.11% to -0.13%).
The researchers report that this is equivalent to a reduction in six preterm births per 1,000 deliveries over the five years following the second phase of the ban.
The researchers concluded that there were “significant reductions in the rate of preterm births after the implementation of different types of smoking bans, whereas no such decrease was evident in the years or months before these bans” and that this has important public health implications, given the association between preterm birth and the baby’s health.
This study suggests that the rate of preterm births dropped in the years immediately after a public smoking ban was introduced in Belgium. This is not to say that the ban was the sole factor contributing to a change in the risk of preterm birth.
The study authors suggest that their research is best viewed and interpreted as: “an investigation into the possible impact of a ‘population intervention’ rather than an investigation of changes in individual behaviour”. They suggest that the trend in preterm births that they observed could possibly be due to the impact of unmeasured confounding variables, and not to the smoking ban.
They note that other outcomes were measured, including birth weight and size for gestational age. No trend over time was observed in these outcomes, despite the fact that they have been previously found to be associated with secondhand smoke exposure.
Given the limitations of a single time-trend study, it is not possible to state conclusively that population-wide smoking bans are associated with reduced risk of preterm birth.
The researchers also note that similar studies in different countries could be useful in determining whether this trend is consistently seen after smoking bans are introduced, and whether reverse trends are seen in countries in which such bans have been introduced but later relaxed. Of course, we would like to see the results for similar research in this country.
Despite these inherent limitations in interpreting the results of this study, it is still the case that smokers should avoid smoking near pregnant women and that pregnant women should avoid smoking and smoky environments.