Establishing Methods to Determine Clinically Relevant Bolus and Basal Rate Delivery Accuracy of Insulin Pumps

Abstract

This article proposes an approach on how to test, evaluate, and present bolus and basal rate accuracy of insulin pumps from a clinical perspective.

Accuracy of Bolus and Basal Rate Delivery of Different Insulin Pump Systems

“…To our knowledge, this is the largest comparison of insulin pumps tested in parallel published so far, the used methods were described and discussed in detail before. 9 Factors such as the so-called siphon effect, 11 infusion set changes, 12 or pressure variations, which might have influenced the results or alleviated the inherent differences between the pumps, were avoided as far as possible. In practice, the tested insulin pumps might not be equally sensitive to these factors.…”

“…The latter has previously been intensively discussed and was, therefore, addressed separately by us. 9 In this study, accuracy of bolus and basal rate delivery was tested in vitro based on IEC 60601-2-24 8 for different insulin pump systems available in Germany at the time of testing.…”

“…Test setups and procedures were described in detail before. 9 In short, accuracy was determined by means of weight differences through insulin delivery into a water-filled oilcovered beaker placed on a balance as described in IEC 60601-2-24. 8 For the patch pump, the setup had to be modified due to the lack of an administration set.…”

“…8 For the patch pump, the setup had to be modified due to the lack of an administration set. 9 Bolus accuracy was determined for a standard bolus of 1 and 10 U. For the 1 U bolus accuracy, 25 boluses were delivered and weighed individually.…”

“…The 72 h of basal rate delivery was divided into 1-h windows and deviation from target (1.0 U/h) was calculated for each 1-h window, and the percentage of 1-h windows within different ranges from target (-15%, -10%, -5%) was calculated. 9 In addition, the total deviation for the whole 72 h and for the first 12 h was determined, respectively. Trumpet curves as required by IEC 60601-2-24 were generated by calculating the minimal and maximal deviation for observation windows of 15, 60, 150, 330, 570, and 930 min, respectively, whereas the first 24 h were neglected.…”