Preoperative risk prediction tools that predict morbidity risk in adults undergoing surgery: An Evidence Review

Risk prediction tools play a critical role in preoperative care by estimating the likelihood of adverse outcomes, including mortality, morbidity, and postoperative complications. In low-risk surgical settings such as surgical hubs, accurate risk prediction is particularly valuable. The aim of this review was to identify and map the evidence for 14 validated pre-operative surgical risk prediction tools currently used in Wales within any elective, or non-emergency surgical setting, and to provide a more in-depth look at the findings for a selection of tools deemed to be the most applicable on a population level to the context of surgical hubs. Included studies were published between 1999 and 2024. No evidence was found for two of the risk prediction tools however, a total of 118 studies were identified across 12 risk prediction tools. None of the evidence found was looking at the predictive ability of risk prediction tools for selecting patients suitable for surgical hubs. The tools were used across a range of surgical specialties and measured composite complications, individual complications, and healthcare utilisation and recovery measures. No risk prediction tool adequately predicted complications across all surgical specialties. Among the included studies, there was considerable heterogeneity in which surgical specialties the risk prediction tools were used for, how complications were defined, and which measures were used to determine a tools predictive ability. This makes direct comparisons very challenging. Four tools were selected as being potentially the most impactful at a population level for a more in-depth look at the findings: ACS NSQIP, P-POSSUM, RCRI, ASA classification system. A total of 76 studies were identified across these 4 tools. Key findings for the four risk prediction tools of interest are described. Overall, no one tool was identified that adequately predicted complications across all surgical specialties. The predictive ability of the tools varied across different surgical specialties. Further research using consistent methods is needed to better understand the predictive ability of risk prediction tools and allow a robust evaluation. Given no single risk prediction tool adequately predicted complications across all surgical specialties, it may be likely that some tools are better suited for specific surgery types or that a combination of risk prediction tools may be needed to adequately assess an individuals level of risk. Funding statementThe authors and their Institutions were funded for this work by the Health and Care Research Wales Evidence Centre, itself funded by Health and Care Research Wales on behalf of Welsh Government.