The information in the present study is related to the prevalence of type 2 diabetes and the incidence of thyroid cancer which was extracted from the Global Burden of Disease (GBD) site during 1990-2019 in countries: Latin America, Australia, Caribbean, Central Asia, Central Europe, Central Latin America, Sub-Saharan Africa, East Asia, Southeast Asia, South Latin America, tropical Latin America, Western Europe, Sub-Saharan Africa, East Sub-Saharan Africa, South Asia, Oceania, North Africa, and the Middle East were the high-income parts of North America, Asia and the Pacific (
19). In order to conduct this study, all the required information related to five continents and based on 204 different countries were extracted from the GBD site, which includes the incidence of thyroid cancer rate (per 100 thousand) during 1990 - 2019 and the prevalence of type 2 diabetes rate (per 100 thousand). All values of the studied indicators were calculated according to the world population, which is estimated by GBD and standardized in terms of age.
3.1. Statistical Analysis
The main objective of a longitudinal study is to determine the change in the response variable over time and the factors affecting the changes (
20,
21). On the other hand, given that in longitudinal data, the assumption of independence of observations is violated, so for data analysis, it is necessary to use methods that can consider this correlation (
22,
23).
In the random effect model for longitudinal data, it is assumed that there is a kind of normal heterogeneity among the subjects in the study population this heterogeneity between the subjects can be considered by considering the random effect in the model. In this case, it is possible to make individual predictions for the random effect model (
21,
24,
25). For this purpose and to investigate the relationship between the prevalence of type 2 diabetes and the incidence of thyroid cancer, a random effect model was used. In this study, in order to accurately estimate the coefficients, a random effect model was used which randomly considered both the intercept and the slope. Here, the random effect of intercept was the initial thyroid cancer unique to each country and the random effect of the slope was the trend of unique changes in cancer incidence attributed to the prevalence of diabetes in each country.
The model used in the present study is as follows:
Where is the occurrence of thyroid cancer in the country i (i = 0,…, 204) at the time j (j = 0,…, 30). In the above model, and are constant coefficients of intercept and the slope of the regression line, respectively, and and are the random effect of intercept and the slope of the regression line of the model, respectively. is also a random error in the country i at the time j, the distribution of all these random effects follows the normal distribution.
In the present study, in order to determine the relationship between the incidence of thyroid cancer and the prevalence of type 2 diabetes using the random effects model, “nlme” package in R software was used.
This study was approved with the reference number of IR.SSU.SPH.REC.1399.090 by the Ethics Committee of Shahid Sadoughi University of Medical Sciences, Yazd, Iran.