The Effect of Frequency of Technology Uploads on Glycemic Control in Adult Insulin Pump Users

Dear Editor,

In an effort to deliver insulin in a more physiological manner, insulin pumps and continuous glucose monitors (CGMs) have been developed.¹ Insulin pumps are more widely used than CGMs, but in some cases, the two are integrated into a single system. This integrated system has shown efficacy in reducing hypoglycemia² and improving quality of life.³ Even so, many users fail to achieve their target glucose levels.

An additional utility that comes with some insulin pumps is the ability to store and track clinical data, namely, the user’s blood glucose levels. For Medtronic insulin pumps, this is accessed through the Carelink website. It includes the capability to chart a patient’s fingerstick, insulin pump, and, if applicable, CGM data. This enables clinicians and patients to gain a greater grasp of their patients’ daily blood glucose patterns and to make more well-informed recommendations for their insulin dosing. The data must be manually uploaded by patients, and the frequency of uploads varies heavily between individuals. Patients attending the Centre of Diabetes, Endocrinology, and Metabolism at Galway University Hospital are encouraged to upload regularly and before every appointment, but this is not always achieved because of a lack of computer access and behavioral factors. We hypothesize that an increased upload frequency is a surrogate for getting the most out of the device and data and would correlate with the magnitude of improvement in glycemic control if any.

Through accessing patient records, with consent, glycated hemoglobin (HbA1c) values before commencing insulin pump therapy were retrieved. The difference between this and their most recent value can be seen as a measure of improvement of their glycemic control; the lower the value, the greater the improvement. By accessing the patients’ data through the Carelink online resource, the dates of their five most recent uploads can be seen, and the mean number of weeks between uploads calculated; a lower number suggesting a higher frequency of uploads. This was done for 20 patients (10 males and 10 females). A Pearson correlation between the change in HbA1c and the mean number of weeks generated a value of 0.44 (P = .02). This shows a mild positive correlation. Put another way, those with greater gaps between uploads tended to have little to no improvement, or even deterioration, of their HbA1c. This data can be seen on a scatter plot (Figure 1).

Figure 1.

A scatter plot illustrating the relationship between the average number of weeks between uploads and the change in HbA1c over the course of insulin pump therapy. N = 20. Line of best fit shows a moderate, positive correlation between the two.

It remains to be seen whether there exists any causality between these two variables. It seems likely that patients more involved and educated in their treatment would be more likely to frequently upload their data, as well as have better outcomes. Even if there exists no direct link, it remains plausible that encouraging frequent uploading would also encourage patients to take a more active role in their treatment, leading to better outcomes.