Closed‐Loop Insulin Delivery in Inpatient Care: A Systematic Review and Meta‐Analysis of Clinical and Implementation Outcomes

Abstract

ABSTRACT Aims Closed‐loop (CL) insulin delivery systems have transformed outpatient diabetes care. However, evidence supporting their role in inpatient settings is limited. This systematic review and meta‐analysis evaluated the safety, efficacy, and implementation of CL systems in inpatient care. Materials and Methods This review followed PRISMA guidelines (PROSPERO: CRD420251133486). Eligible studies reported glycaemic, safety, and/or implementation outcomes for adults managed with CL systems in inpatient care. EMBASE, PubMed, MEDLINE, Scopus, CINAHL, and Cochrane Library were searched from inception to August 2026. Risk of bias was assessed using RoB2 and ROBINS‐I. Random‐effects meta‐analysis evaluated glycaemic outcomes and length of stay (LOS). Results were presented as pooled mean differences (MDs) and 95% confidence intervals (CIs). Implementation data were analysed thematically. Intrapartum outcomes were analysed separately. Results Thirteen inpatient studies were included, nine contributing to meta‐analysis. CL delivery increased time in range (MD +22.27%, 95% CI: 16.03–28.51; p < 0.001) and reduced time above range (MD −23.75%, 95% CI: −28.05 to −19.45; p < 0.001), mean glucose (MD −1.66 mmol/L, 95% CI: −2.29 to −1.03; p < 0.001) and glucose standard deviation (MD −0.64 mmol/L, 95% CI: −0.85 to −0.43; p < 0.001). No significant differences were observed in time below range, coefficient of variation or LOS. Total daily insulin dose was higher (MD +10.82 IU/day, 95% CI: 0.02–21.62; p = 0.05). No serious device‐attributed adverse events were reported. Implementation challenges included staff training, technical support, and workflow integration. Four additional intrapartum studies were included; two contributed to the meta‐analysis, showing increased pregnancy‐specific time in range (MD +11.82%, 95% CI: 5.56–18.08; p < 0.001). Conclusion CL systems improve inpatient glycaemic control, although effects on LOS and comparative safety remain uncertain. Wider adoption requires sustainable technical and organisational support.