Background: The defining characteristic of a longitudinal study is that subjects are measured repeatedly through time. Longitudinal studies are in contrast to cross-sectional studies, in which a single outcome is measured for each individual. The primary objective of this study is to use copulas to model the within-subject dependencies over time.

 Methods: In this longitudinal study, we used the hospital records of patients with type 2 diabetes in Lolagar hospital, Tehran. Information of patients who visited the hospital at least twice during the years 2006-11 were recorded. Factors affecting fasting blood sugar were determined by regression model and the use of copula functions. We used the Residuals pp plot of copula function for selecting copula. Fitting model was done with R software.

Results: In this study, only three explanatory variables were statistically significant. Smoking (p<0.001), family history (p<0.001), and duration of illness (p<0.001) were the positively significant variables. The coefficient estimate of duration of illness was 0.003, meaning that other variables remained the same, as duration of illness increases by one unit, the expected value of fasting blood sugar will increase. Conclusion: In addition to identifying risk factors of fasting blood sugar, it was shown that use of copula function is an appropriate method for longitudinal data analysis and modeling correlations between data.