|
A patient with significant clinical complexity with prior history of hypoglycemic events (home regimen included multiple daily injections of insulin) was admitted with hyperglycemia (>300 mg/dl) on [admission date unknown to glytec].On (b)(6) 2026, the provider initiated glucommander using weight-based dosing, (the patient's weight was 58.9 kg), resulting in initial orders of 15 units basal daily and 5 units lispro per meal.The initial basal and prandial doses were determined by the provider and the orders were entered in the ehr.Patient received 1 unit correction at initiation and the 15-unit basal as ordered; patient did not receive any mealtime insulin.Within approximately 12-18 hours, patient experienced severe hypoglycemia, required resuscitative treatment, and experienced recurrent hypoglycemic episodes.Patient expired approximately 30 hours after the initial dosing event.Customer noted that the patient's death was not attributed solely to hypoglycemia but occurred during a critical illness window.
|
|
Glytec's investigation included audit log review, communication with the customer, and analysis of information provided by the user facility.The audit log confirmed that glucommander's algorithm operated as designed and in accordance with its intended use.No software malfunction, algorithm error, or device deficiency was identified.From the collaborative case review glytec facilitated with the customer, glytec learned that the initial basal dose of 15 units was derived from provider-entered weight-based dosing.The customer's order set was configured with a multiplier of 0.5 (glytec's best practice recommendation for high-risk renal/dialysis patients is a multiplier of less than or equal to 0.3).The provider had limited awareness of available conservative dosing options within their ehr, including the custom dosing feature, the inpatient calculator, and multiplier selection guidance for this patient population.Paraphrasing a statement from one of the use facility's pharmacists made during the case review: "from a dosing standpoint, everything appeared appropriate when compared to the patient's home insulin regimen".The patient's usual total daily dose was around 24 units, and given that their blood glucose levels were over 300 at admission, a basal dose of 15 units of lantus was considered reasonable from the pharmacy's perspective".Glytec's investigation determined that at the time of the adverse event, glucommander had not generated any algorithm-derived dosing recommendations.The doses administered to the patient, 15 units basal and 1 unit correction, were based entirely on provider-entered orders.Glucommander received and displayed these provider orders as designed, but had not yet received sufficient post-initiation patient data from the ehr to generate its own recommendations.The samd's intended use (iterative, algorithm-driven insulin dosing guidance based on ongoing blood glucose data and patient response) was never realized in this case, as the patient experienced the adverse event prior to any algorithm-generated recommendations being produced.Note: glytec requested confirmation of date and cause of death from the reporting facility.No response was received at the time of this report.Following completion of its investigation, glytec provided best practice recommendations to the customer following this event, including clinical education reinforcement, development of custom provider training materials, and recommendations regarding order set configuration.Glytec recommends that providers consider more conservative starting doses, (or the home regimen, if known), when appropriate at the time of ordering.Implementation of order set configuration and clinical workflow decisions remains within the customer's clinical authority.
|
