The National Institute for Health and Care Excellence (NICE) guideline for self-harm advises against the use of risk assessment tools to predict future occurrence of repeat self-harm or suicide in individuals who have self-harmed, or to inform decisions regarding their treatment and discharge. In this perspective article, we discuss shortcomings in the process of developing this guideline, including: (1) limitations in the NICE evidence review underpinning these recommendations, which resulted in very minimal evidence being included; (2) developing definitive recommendations and drawing strong conclusions regarding the limited predictive ability and potential harms of tools, which were almost entirely based on the committee's expertise and experience and (3) not acknowledging the uncertainty and gaps in the evidence base, particularly around model impact, acceptability and feasibility. We highlight new evidence since this 2022 guideline, including examples of international work assessing model implementation and cost-effectiveness. We propose that there is an urgent need for more rigorous primary research assessing model impact, feasibility and acceptability, as well as empirical work addressing concerns about potential harms and misuse of tools, notably the denial of care. While prediction models should not be prematurely implemented in clinical practice without adequate validation and impact assessment, well-developed and validated tools in this area have the potential to improve clinical care for individuals who self-harm. Future updates to the guideline should be informed by emerging higher quality evidence in the field.