Compliance to self-monitoring of blood glucose: a marked-item technique compared with self-report

This study compared subjects' self-reported rates of compliance to self-monitoring of blood glucose (SMBG) with an objective measure based on a "marked-item" technique. We followed 25 obese patients with type II diabetes who were participating in a behavorial weight control program and monitoring their blood glucose with Chemstrips bG (Bio-Dynamics, Inc., Indianapolis, Indiana). Subjects' self-report significantly overestimated actual compliance as assessed by the marked-item technique. Moreover, the self-report measure failed to identify 35-45% of the noncompliant patients. Compliance decreased steadily over the course of the 37-wk program. Accuracy of SMBG was less problematic than compliance; 85% of patients were able to read Chemstrips bG within 20% of actual blood sugar, and the average blood sugar reading obtained from 2 mo of SMBG correlated highly (r = 0.78, P less than 0.01) with HbA1. Our data suggest that objective measures such as the marked-item technique described in this article should be used to assess compliance to SMBG and behavioral strategies to improve compliance should be developed.