Childbirth, the extension of the family, has a prominent place in communities and cultures around the world. Yet, in modern society, the actual setting of childbirth has become the hospital—that detached place of ‘sickness repair’—rather than the community and family environment. This has for many years sustained a debate in obstetrics, about the wisdom of this ‘medicalisation’ of childbirth. Medical professionals including general practitioners, midwives and obstetricians and increasingly women as the prime stake holders of ‘their’ delivery, have questioned why in healthier than ever societies, the hospital should be the host environment for the significant event of the birth of a new child. The discussion is a mixture of ideology—‘home is best’ versus ‘hospital is safe’—and of clinical evidence of the risks related to a community-based delivery, and on the ability of the hospital setting to reduce this risk. In June’s issue of the journal, two papers that contribute evidence to the debate are published, a critical appraisal of the outcome of intentional and unintentional home births in the UK, and an analysis of primary care midwives’ deliveries in the Netherlands. Three themes develop from these papers that warrant further discussion: first, the quality of evidence on place of birth; secondly, the effectiveness of risk selection in primary midwifery care and thirdly the need to give women a truly informed choice. The gist of the papers is that there is insufficient evidence of an elevated risk of home deliveries, and consequently there are no grounds to discourage healthy, low-risk women from giving birth at home. Gyte et al. present a critical appraisal of an earlier study of intentional and unintentional home deliveries in the UK. They correctly argue for the appropriate denominator—planned home deliveries—in calculating adverse outcomes of home births and make a convincing case that the number of intrapartum related neonatal deaths is in all probability lower than earlier reported. But this does not answer the question of ‘how low an intrapartum related neonatal death rate is good enough’ And more disturbing is their overall conclusion that no reliable data are available for the UK, and that the prevailing guidelines may be based on wrong assumptions. Given the demand for home birth, this is an unsatisfactory state of affairs. The first issue to consider is the place of delivery. This brings us to the second paper Amelink-Verburg et al., which shows that there is more at stake than ‘home versus hospital’. In the Dutch healthcare system, community-based midwives play a leading role in the delivery of low-risk women and they operate as independent professionals in primary care. Their domains of supervising deliveries are the family’s home and also the primary care obstetric facilities in the hospital. This is a different perspective, in particular with respect to one of the main ‘outcomes’, intrapartum referral. In this context, an important indicator is the ability of early intrapartum referral, allowing nonurgent transfer to the hospital. Here, the ‘primary care’ function of the midwife becomes evident. An essential feature of primary care is the assessment of risk as early as possible in the episode. From this point, the most appropriate level of care is determined (self-care; professional care within the primary care team; or referral to a specialist). This ensures access to care for all (equality), while focusing care to those with the highest needs (equity), and guarantees cost-effectiveness. The ‘Dutch’ approach to birth at home follows from this approach. The quality of risk assessment is vital in this context, in this case to help the pregnant woman to make an informed choice for her place of delivery. The analysis of Dutch midwifery practice gives some insight into the risk selection in primary obstetric care. This risk-selection is built on consensus between midwives, obstetricians and general practitioners. In fact, this consensus between obstetricians and general practitioners has much deeper roots and