First, we question the underlying premise of this debate: that hospital birth is inherently safe. There has been much discussion regarding the ‘cascade of interventions’ that has become ubiquitous, at least in US hospital settings, and which culminates in caesareans—many of which are medically unnecessary—for over 30% of labouring women. Caesareans result in substantial morbidity, both in the current and subsequent pregnancies, and interventions drive up the cost of birth to unsustainable levels. This is the standard towards which we should strive? An examination of the studies cited by those on each side of this debate reveals an interesting pattern: proponents and opponents are relying on two distinct types of data sources. Proponents tend to cite studies that use data from perinatal registries, which capture both intended and actual places of birth. Registry-based research, much of it population-based, from Canada, the UK (Birthplace in England Collaborative Group BMJ 2011;343:d7400), Norway, the Netherlands (de Jonge et al. BJOG 2015;122:720–728), and the US (Cheyney et al. J Midwifery Womens Health 2014;59:17–27) all report similar outcomes for low-risk women. Homebirth in these studies is consistently associated with high rates of normal physiologic birth (caesarean rates <10%), low morbidity, reduced healthcare costs, low intrapartum and neonatal death rates but with, unsurprisingly, higher intrapartum mortality for higher-risk births (Cheyney et al., as above). Opponents tend to cite studies based on US vital statistics data. These studies (Grunebaum et al. Am J Obstet Gynecol 2014;211: 390e1–7), and one controversial meta-analysis (Wax et al. Am J Obstet Gynecol 2010;203: 243.e1–8), argue that homebirth is associated with elevated neonatal death rates. Unfortunately, birth certificates in the USA do not universally capture planned place of birth, which leads to misclassification bias when conducting analyses that require an intention-to-treat model. Why, then, is there still a debate about the basic safety question? Homebirth clearly can be as safe (and for some outcomes, safer) than hospital birth (de Jonge et al., as above). Registry-based research, conducted in multiple cultural settings, medical systems, and with enormous variability in training of midwives, supports this conclusion. Certainly it is safe in settings where homebirth midwives are well integrated into the overall healthcare system, as evidenced by the extremely low intrapartum mortality rates achieved in the Netherlands (de Jonge et al., as above). The questions we should be asking instead are: ‘Safe for whom?’ and ‘Under which circumstances?’ What the balance of evidence suggests is that homebirth safety may be contingent upon getting certain things right: the patient population, the collaborative relationships between midwives and physicians, and the transfer criteria. These contingent factors are not yet in place everywhere, including in the US. Yet, the strength of the Netherlands data is that it demonstrates that homebirth can be safe; where it is not, this should prompt critical reform of practice, not retreat from homebirth as an option. Finally, although this conversation has been dominated by discourses of risk and danger, homebirth is also clearly associated with less frequently discussed benefits: high rates of breastfeeding, normal physiologic birth, and patient satisfaction. Let us move beyond this polarising debate on whether homebirth is safe, towards a more nuanced and collaborative research agenda aimed at improving access to high-quality maternity care for all women, across all settings. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.