Assessing the quality of glucose monitor studies: a critical evaluation of published reports.

BACKGROUND In recent years, a large number of studies have been published on the performance of glucose monitors. The quality of these reports is not known. METHODS We searched the PubMed database for performance evaluations of handheld glucose monitors published from August 2002 to November 2006. Relevant articles were compared to 20 recommendations from the Standards for Reporting Diagnostic Accuracy (STARD) and 18 recommendations from the Clinical and Laboratory Standards Institute (CLSI). RESULTS A total of 52 reports met our inclusion criteria and were reviewed. None (0%) of the reports conformed to all 38 STARD and CLSI recommendations. The range of compliance to these recommendations varied widely (median 53%; range 21%-84%). Only 1 study of the 52 reported following a CLSI recommendation for checking reference test results. Fewer than half (42%) of the reports contained STARD-recommended statements regarding how and when comparative measurements were performed. CONCLUSIONS None of the glucose monitor reports from our review conformed to all STARD and CLSI recommendations. Our finding that the average rate of compliance to recommendations was low suggests that many of the researchers did not follow published recommendations for study design, methodology, and reporting and that study quality and conclusions may have been affected. Future studies evaluating the performance of glucose monitoring systems should be carefully designed and follow published recommendations for methodological and reporting quality.