Domestic Homicide Reviews (DHRs) are a statutory review process to better understand domestic homicide in England and Wales. As a policy intervention, DHRs are intended to build a picture of the circumstances before such deaths and identify gaps in practice, policy and system response. The rationale is that this learning can improve response to domestic violence and abuse and reduce the likelihood of future homicides. However, little is known about how the DHR process operates, including how knowledge is produced or its subsequent use, including any outcomes. In effect, for the most part, DHRs are a ‘black box’. Yet, researchers are increasingly using DHR reports as a source of data. By locating ourselves within these processes, this article explores the implications of limited engagement with DHRs as a process of knowledge generation to date. It focuses on the implications for researchers, in particular the epistemological and methodological issues that arise, before considering what this might mean for policy and practice. It identifies recommendations to address key gaps in the understanding and use of DHRs for research purposes.<br/><br/>Key messages<br/><br/>Recognise the potential and challenges of using DHR reports as data.<br/><br/>Consider the everyday work processes associated with the production of DHR reports.<br/><br/>Concepts in DHRs must be clearly defined to enable robust data collection.<br/><br/>Develop a feedback loop between research and practice so each can benefit from and inform the other.