Development of a cross-disciplinary continuous insulin infusion protocol for non-critically ill patients in a French university hospital.

RATIONALE, AIMS AND OBJECTIVES:In non-critically ill patients, the use of an insulin syringe pump allows the management of temporary situations during which other therapies cannot be used (failure of subcutaneous injections, awaiting advice from the diabetes team, emergency situations, prolonged corticosteroid therapy, initiation of an artificial nutrition, need for a fasting status, etc.). To manage the risks related to this «never event», the use of a standard validated protocol for insulin administration and monitoring is an essential prerequisite. To this end, a multidisciplinary approach is recommended. METHOD:With the support of our subcommission «Endocrinology-Diabetology», we proceeded with a «step-by-step process» to create such a standardized protocol: (1) review of all existing protocols in our hospital; (2) overview of the literature data concerning insulin infusion protocols developed by multidisciplinary teams in France and abroad; (3) development of a standardized protocol for non-intensive care unit patients, respecting the current recommendations and adapting it to the working habits of health teams; and (4) validation of the protocol RESULTS:Two protocols based on the same structure but adapted to the health status of the patient have been developed. The protocols are divided in to three parts: (1) golden rules to make the use of the protocol appropriate and safe; (2) the algorithm (a double entry table) corresponding to a dynamic adaptation of insulin doses, clearly defining the target and the 'at risk situations'; and (3) practical aspects of the protocol: preparation of the syringe, treatment initiation and traceability. The protocols have been validated by the institution. CONCLUSION:Our standardized insulin infusion protocol is simple, easy to implement, safe and is likely to be applicable in diverse care units. However, the efficiency, safety and the workability of our protocols have to be clinically evaluated.