Nutrition delivery, workload and performance in a model-based ICU glycaemic control system

Area of Science:

• Critical Care Medicine
• Endocrinology
• Clinical Nutrition

Background:

• Hyperglycaemia is common in adult intensive care units (ICUs), linked to increased morbidity and mortality.
• Effective glycaemic control (GC) is challenging but crucial for reducing ICU patient morbidity and mortality.
• The STAR protocol optimizes GC by adjusting both insulin and nutrition, but its variable nutrition component is less clinically desirable.

Purpose of the Study:

• To compare the performance of the STAR insulin protocol with three simpler, alternative nutrition protocols.
• To evaluate GC performance, safety, and intervention workload in virtual trials.
• To identify a more clinically acceptable nutrition protocol for use with the STAR GC system.

Main Methods:

• Simulated STAR insulin protocol with fixed, CB stepped, and SLQ stepped nutrition protocols in a 221-patient virtual cohort.
• Assessed glycaemic control (GC) performance, safety (severe hypoglycaemia), and total intervention workload.
• Compared outcomes against the current variable nutrition protocol used with STAR.

Main Results:

• All alternative nutrition protocols reduced total intervention workload by 14.6-19.8% through fewer nutrition changes.
• Stepped nutrition protocols (CB and SLQ) achieved GC performance comparable to the current variable nutrition protocol.
• The SLQ nutrition protocol significantly improved GC safety, nearly halving severe hypoglycaemic events (5 vs. 9).

Conclusions:

• The SLQ nutrition protocol is a superior alternative to the current variable nutrition protocol for the STAR GC system.
• Both stepped nutrition protocols can be adapted by STAR to reduce workload and enhance clinical acceptability.
• Adapting stepped nutrition protocols maintains STAR's proven GC performance and safety while improving usability.