Why absence of evidence of risk is not the same as evidence for absence of risk

Yesterday’s news indicating the potential overplaying of the risk of alcohol in pregnancy highlights the professional and ethical tensions that scientists and evidence-based practitioners face on a daily basis. On the one side is the often incredible power of medical and scientific knowledge, on the other the daily encroachment of medicine and science into every aspect of our lives, what is termed ‘medicalisation’.

Many academics and even practitioners have presented coherent and powerful arguments against medicalisation, often identifying the worrying role of commercial, political and similar interested agencies in its apparently insidious spread.

But there are occasions in which these arguments, no matter how well made, no matter how reasonable, and no matter how well intentioned, can be counter-productive. Yesterday’s media around alcohol and pregnancy may be just such an example.

As scientists with a keen interest in exercise and pregnancy, we have to daily consider the tensions between common sense and an un-medicalised approach, and one based on scientific evidence, even if that evidence is not complete. The bottom line, however, is that we also have to veer on the side of caution.

Exercise was once considered a risk to pregnant mothers, we now know the benefits. The risks were overplayed in early research, and subsequent evidence indicated the positive effects of exercise on health during pregnancy. Likewise, evidence for the harmful effects of alcohol may have been overplayed, but where is the evidence for benefit? It is a maxim of knowledge that lack of evidence for something is not the same as evidence for the lack of something. Unlike many areas of science, it is problematic to research the real effects of alcohol during pregnancy, it is an emotive issue riddled with ethical and pragmatic challenges.

But there is a further risk. Any practitioner in public health will be happy to tell you that many people will hear what they want to hear, they will ignore the 99% of messages that indicate, for example, that smoking is harmful, and hear the 1% that indicates a lack of harm.

Scientists and academics should challenge scientific data, it is how science and knowledge progress. But in doing so we must exercise caution because the difference between the message sent – the evidence is not as strong as we like to think – and the messages that people hear – that it’s not as dangerous as we thought to drink alcohol during pregnancy – is stark.