Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

© 2020 Background and objectives: We aim to evaluate the reliability and compliance of self-monitoring of blood glucose (BG) measurements in women with gestational diabetes (GDM). We developed an automated way of comparing self-reported logbooks BG measurements and those held on electronic glucometers. Existing techniques involve manual inspection of the data, an approach that is inefficient, potentially inaccurate and not scalable. To enable analysis of large datasets, we employ an algorithmic approach that automatically matches measurements from both sources to compute their concordance. Methods: The algorithm takes into account missing samples, time mismatches and varying measurement schedules. We applied this strategy to a dataset of 102 participants with gestational diabetes mellitus (GDM) who were the control arm of the TREAT-GDM study. After being diagnosed with GDM, women were given an electronic glucometer and were instructed to report their glucose on a paper logbook up to 6 times a day for the remainder of their pregnancy. At the end of the study, logbooks were manually digitised and glucometers' memories were downloaded. A set of metrics related to the reliability of logbooks and patients’ compliance to the measurement routine were computed by our algorithm. Results: In total, we analysed 8179 logbooks and 9511 m entries. To take into account differences in the time-stamps of the two data sources, we show that a tolerance of 90 min has to be given, which increases to 180 min if we also want to match logbook entries where no time was specified. In terms of reliability, among the reported measurements, 70% were concordant with the records on the glucometers. In terms of compliance, participants reported 6 measurements per day in 47% of the days they were supposed to. Conclusions: This paper describes an algorithm for automatic comparison of logbooks and glucometers' records that allows the analysis of the reliability and compliance at scale. In accordance to other studies including patients with GDM, our cohort confirms that women's compliance to blood glucose monitoring is low. Our analysis shows that problems due to the reliance on paper records, e.g. data-loss, can be mitigated by automated analysis of electronic records.

Original publication

DOI

10.1016/j.imu.2020.100397

Type

Journal article

Journal

Informatics in Medicine Unlocked

Publication Date

01/01/2020

Volume

20