998-P: Real-World Evidence for Computerized Insulin Dose Adjustment Algorithms (CIDAAs) in the Effective Use of Continuous Glucose Monitoring (CGM) by Primary Care Clinicians (PCCs)

Introduction & Objective: PCCs are challenged using insulin because of time constraints and experience. Mellitus Health has developed FDA-cleared CIDDAs that enabled PCCs to effectively lower A1C levels by 1.5-2.0% over 3-6 months in poorly controlled, insulin-requiring patients who provided remote glucose readings to HIPAA approved cloud-based CDAAs that analyzed them and sent dose change recommendations to PCCs. Inconvenience and discomfort limit glucose fingerstick tests, barriers which CGM overcome. This study examined the ability of these CIDAAs to help PCCs effectively utilize CGM in similar patients.

Methods: A clinical pharmacist used a Free Style Libre Pro in 23 patients and an NP used a Dexcom reader in 13 patients, both in real world situations. Since A1C levels were not scheduled at specific times, the outcomes were changes in average glucose (AG) levels and calculated eA1Cs between the 1^st and last reports sent to the PCC who could accept or modify the dose adjustment recommendations. The new doses served as the basis for the next report.

Results: A mean of 159 reports were issued over a mean of 113 days. Initial and final insulin regimens were basal alone, 15 and 12, basal/bolus, 20 and 23, and 1 short-acting alone in both. Baseline AGs (mg/dL ± SD) fell from 230 ± 68 to 196 ± 56 and eA1Cs (% ± SD) from 9.6 ± 2.4 to 8.4 ± 1.9 (P <0.001 by paired t tests). The PCCs accepted 56% of the recommended changes. Percents of glucose values <70 mg/dL were <2.0% and not significantly different between 1^st and last reports signifying the safety of these CIDAAs. The total doses of insulin increased by 43%.

Conclusions: If a higher % of recommendations had been accepted, there may have been greater decreases. CIDAAs can facilitate and educate PCCs in the effective use of CGM.