Diabetes Mellitus Type 2 (T2DM) – Prevalence Differentials in the US and in Europe: Reality or Measurement Error?

Introduction: Cancer, Metabolic Syndrome, Dementias are the dominant health problems in aging societies, T2DM advancing all three. There are considerable regional differences in T2DM prevalence which may be real or not. Today most new cases in North America and in Europe are found in screening tests in high risk individuals which still do not present clinical symptoms of T2DM, a substantial proportion of the true prevalence of T2DM has to be estimated. There is a wide range of approaches for estimating undiagnosed T2DM, none of them established so far. The OECD multiplied observed prevalence for the UK by a factor of 1,5 and doubled for other European countries. Novel approaches in Germany estimate undiagnosed diabetes prevalence as of 25 – 30% of he diagnosed cases, and way below the 100% suggested by the OECD. In the US the lowest total estimated prevalence of diabetes is found in Midwest and Northeast, the highest in Southern and Appalachian states, with a variation of app. 150% as compared to the lowest. In the European Union, prevalence variation for ages 20 – 79 is even higher at 180 – 200% highest in Portugal, Germany vs. UK, Sweden. Data and Methods: Obesity, mortality from coronary heart disease (CHD) and stroke, can be determined straightforwardly – and all three are closely related with T2DM. This is a suitable situation where to apply an ecological approach. Results: Among the 50 US states we get ordinal correlation coefficients between adipositas and diabetes prevalence at ages ages 60+ of Kendals tau_b 0.328 at p = 0.001, and Gamma = 0.333 at = 0.001. Among the 20 European States for which the OECD has data we get ordinal correlation coefficients between adipositas and diabetes prevalence at ages 20 – 79 of Kendals tau_b = 0.265 at p<.150, Gamma = 0.266 at p < 0.150. Furthermore, for the US we get ordinal correlation coefficients between diabetes and CHD mortality at ages 60+ of Kendals tau_b 0.333 at p = 0.001, Gamma = 0.334 at = 0.001 and for stroke of Kendals tau_b 0.622 at p = 0.0004, Gamma = 0.626 at = 0.0004. For the European states, here, we get less conclusive associations: no correlation between diabetes prevalence and CHD mortality, but stroke mortality of Kendals tau_b = 0.302 at p < 0.012, Gamma = 0.305 at < 0.012. Graphical representation of these associations will be shown. Discussion: The results from the US are more relevant, because regional inequalities in health care, especially in acute situations may be smaller than among European states. Region specific influences may play a significant role in the genesis of adipositas as well as diabetes and, therefore, effective prevention strategy could be and should be region specific. This study also is an example for the potential of the ecological design for checking estimations of the population prevalence of one difficult-to-observe parameter using another easy-to-observe parameter, when a causal relation between the two parameters has been established before.