Application of diabetes risk scores in health checkups

Peer-reviewed | Manuscript received: February 26, 2018 | Revision accepted: June 21, 2018

A comparison of the German Diabetes Risk Score (GDRS) and FINDRISK test

Introduction

In the past years, a large amount of prediction models was published enabling the prediction of future development of type 2 diabetes based on known risk factors [1]. Many of these prediction models are based on solely non-invasively assessable risk factors [2] and therefore, seem to be well suitable for application in clinical practice. However, until now implementation of such prediction models in screening was limited [3]. For early detection of type 2 diabetes in Germany, the German Diabetes Association (DDG) already recommends the application of non-invasive tests [4], but there might still be some barriers for the use in clinical practice. For Germany, however, two tests were developed which are also increasingly applied by the general population.

This is on the one hand the German Diabetes Risk Score (GDRS) which was developed at the German Institute of Human Nutrition (DIfE) and was validated in various external populations [5], and the FINDRISK (“find your risk”), which is an adaptation of the Finnish risk score (FINDRISK, Finnish Diabetes Risk Score) [7]. For both tests simple questionnaire versions were developed [6, 8] including modifiable lifestyle factors such as physical activity, diet, smoking behavior or waist circumference as well as non-modifiable risk factors such as age, family history of diabetes or prevalent hypertension. Detailed information regarding the development, calculation and validation can be found in respective previous studies [5–8].

With regard to application of the GDRS [6] and FINDRISK [7], previous cross-sectional studies [9–12] showed that both tests seem to be valid for the identification of type 2 diabetes.

Abstract

In the study “Praxistest Diabetes” (PraDi) we investigated the application of the German Diabetes Risk Score (GDRS) and FINDRISK in health checkups in terms of the predictive ability to identify undiagnosed prediabetes and diabetes. 403 men and women aged 35–70 years participated in the PraDi study. The predictive ability of prediabetes/diabetes (fasting plasma glucose FPG ≥ 5.6 mmol/L) or diabetes (FPG ≥ 7.0 mmol/L) was evaluated with discrimination (area under the receiver-operating-characteristic curve [ROC-AUC]).

With regard to blood glucose tests, 93 participants had prediabetes/diabetes and 7 participants had undiagnosed diabetes. The ROC-AUC for the GDRS for identification of prediabetes/diabetes was 0.78 (95% confidence interval [95%-CI]: 0.73–0.94) which was higher than observed for FINDRISK with 0.73 (95%-CI: 0.68–0.79; p = 0.029). The ROC-AUC for identification of diabetes was 0.83 (95%-CI: 0.73–0.94) for the GDRS and 0.80 (95%-CI: 0.61–0.99) for FINDRISK (p = 0.67). In the context of health checkups, the GDRS showed a more precise identification of patients with prediabetes/diabetes or diabetes compared to FINDRISK.

Keywords: type 2 diabetes mellitus, riskscore, predictive accuracy of diabetes risk tests, German Diabetes Risk Score, FINDRISK

Full text PDF (free access)