What strategies are used to select patients for direct admission under acute medicine services? A systematic review of the literature.

Background: Pressures on hospital emergency care services have led to increasing interest in new models of acute care provision. One such model is a medical emergency department where medical patients are triaged directly to acute internal medicine, without assessment by emergency medicine. The evidence for this model of care is unclear. Design: Systematic review. Methods: Studies included direct referral pathways to acute internal medicine. The protocol was registered prospectively (Prospero: CRD42023495786). Databases searched included MEDLINE (Ovid), The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE in process, Web of Science, CINAHL, and Embase. Studies had no time or language restrictions. Studies were selected based on inclusion and exclusion criteria, assessed by at least two independent researchers. ROBINS I risk of bias assessment was applied to the selected studies and a narrative synthesis was performed. Results: From 4405 abstracts, 89 full text articles were screened and 4 were selected for data extraction. Two studies assessed tools to predict the need for a medical admission and two studies assessed the impact of direct referral pathways to medicine. Risk of bias was mixed, and studies were heterogeneous. However, the studies reported a good ability to appropriately select patients for direct referral to medicine and a reduced length of time to medical assessment. There were no differences in other outcomes such as mortality or overall length of stay. Discussion: The current evidence to support direct admission to medicine, effectively a medical ED, is limited with studies being heterogeneous and of varying quality. Models for patient selection varied, but there was evidence to support accurate, early identification of medical patients and of reduced delays in medical assessment and care. Conclusion: Given these positive early signs of benefit, more studies are needed to design and evaluate care models such as medical EDs. Registration: Prospero Registration Number: CRD42023495786.