8 Prognosis and harm

Obstetricians and gynaecologists often have to assess whether an exposure is causing a harmful effect. The first part of this assessment involves considering how the results were obtained: can the type of study determine a cause-and-effect relationship; were all subjects at similar risk of being exposed to the putative harmful exposure and of developing the outcome of interest; was information about the exposure ascertained in the same way from all subjects; could knowledge of an individual's exposure status have influenced how the outcome was ascertained? Next, the results themselves need to be considered: what was the size and precision of the risk estimate; could there be other explanations, such as bias and confounding, for the results; were there other pointers, like a dose gradient, consistency with other studies or biological plausibility, which suggest a causal relationship? Finally, the clinical implications of the results need to be evaluated: are the findings applicable to today's clinical practice; who do they apply to; what are the numbers needed to harm? By addressing each of these points, clinicians can determine whether an apparently harmful effect is important and so decide whether they need to change their medical practice.