A commitment to quality and safety, together with choice and continuity of care, underpins current directions for the reform of maternity services in the United Kingdom and Australia.1 2 Recognition that midwifery-led care will be central to these changes, with options for women to give birth at home, has stimulated research about the safest and most effective models of care. Symon and colleagues (doi:10.1136/bmj.b2060) compare clinical outcomes between pregnant women accessing an independent midwife and women using the National Health Service (NHS).3 They found that obstetric risk factors were an important predictor of perinatal outcomes. Unwittingly the authors illustrate pitfalls involved in attempting to match disparate datasets. Nevertheless, they confirm some of what we already know about the benefits of continuity models of midwifery care and the important role of risk selection in achieving good perinatal outcomes.

Symon and colleagues attempted to match two groups of women, which is predictably problematic because women choosing independent midwifery care were a self selected group.3 Unfortunately, the matching process was largely unsuccessful, with numerous important differences remaining, including nutritional status, smoking status, alcohol consumption during pregnancy, geographic location of residence, obstetric risk from previous pregnancy, medical complications during current pregnancy, incidence of …