Home birth: gone away, gone astray, and here to stay.

Home birth has attracted a great deal of attention of late, culminating in a meta-analysis to assess its risks for mother and baby. Mothers were estimated to be 2.6 times more likely to die and babies 3 times more likely to die from a planned home birth than from a planned hospital birth. The actual data on which these estimates were based demonstrate that meta-analysis can be developed into an art that suits whatever purpose its authors hope to achieve. Combining studies of home versus hospital, without differentiating what is inside them, where they are, and what is around them, is akin to producing a fruit salad with potatoes, pineapples, and celery.