LINK TO RESEARCH
We sought to examine the correlation between variables and A1C levels to determine if prediction modeling could be used in the screening and diagnosis of diabetes and prediabetes in youth. We also sought to test relationships between A1C levels to insulin sensitivity indices and β-cell function indices. Design and methods: We performed a retrospective review of 904 medical records from youth deemed at-risk for the disease. We performed Pearson correlation, multiple regression, and simple regression testing to determine the relationship between variables and A1C levels. In addition, we performed Pearson correlation testing on insulin sensitivity indices and β-cell function indices to determine the strength of correlation to A1C levels.
Results: Statistical analysis did not show a strong relationship between the variables tested and the A1C. When racial and ethnic groups were tested together, the results from African American participants resulted in bias estimates, and as a result, a statistical model for the entire sample could not be performed. Results indicate that A1C is correlated with all β-cellfunction proxy measurements and correlated to the corrected insulin level at 30minutes, but not the fasting insulin or insulinogenic index.
Discussion: The results from this study underline the multi-dimensional causes of diabetes and prediabetes and further stress the difficulties in predicting the diseases. The causes of diabetes and prediabetes are multifaceted, often individualized, and often difficult to ascertain. Practice implications: Clinicians should continue to examine a variety of variables prior to determining the need for diabetes diagnostic testing.
Interested in this research and working with us? Please contact us here.