Spatiotemporal trends and influence factors of global diabetes prevalence in recent years

Diabetes is one of the most widespread global epidemics and has become the main component of the global disease burden. Based on data regarding the prevalence of diabetes in 203 countries and territories from 2013 to 2017, we employed the Bayesian space-time model to investigate the spatiotemporal trends in the global diabetes prevalence. The factors influencing the diabetes prevalence were assessed by the Bayesian LASSO regression model. We identified 77 (37.9%) hotspots with a higher diabetes prevalence than the global average, 10 (0.4%) warm spots with global average level and 116 (57.1%) cold spots with lower level than global average. Of the 203 countries and territories, 68 (33.5%), including 31 hotspots, 5 warm spots and 32 cold spots, exhibited an increasing trend. Of these, 60 experienced an annual increase of more than 0.25%, and 8 showed an increasing trend. Three populous countries, namely China, the USA and Mexico, exhibited a high prevalence and an increasing trend simultaneously. Three socioeconomic factors, body mass index (BMI), urbanization rate (UR) and gross domestic product per capita (GDP-PC), and PM_2.5 pollution were found to significantly influence the prevalence of diabetes. BMI was the strongest factor; for every 1% increase in BMI, the prevalence of diabetes increased by 2.371% (95% confidence interval (95% CI): 0.957%, 3.890%) in 2013 and by 3.045% (95% CI: 1.803%, 4.397%) in 2015 and 2017. PM_2.5 pollution could be a risk factor, and its influencing magnitude gradually increased as well. With an annual PM_2.5 concentrations increase of 1.0% in a country, the prevalence of diabetes increased by 0.196% (95% CI: 0.020%, 0.356%). The UR, on the other hand, was found to be inversely associated with the prevalence of diabetes; with each UR increase of 1%, the prevalence of diabetes decreased by 0.006% (95% CI: 0.001%, 0.011%).