The use of the Finnish Diabetes Risk score in a mobile application for identifying persons with hyperglycaemia and predicting type 2 diabetes in the Hong Kong population

Abstract

Diabetes is one of the most common chronic diseases globally and is a major cause of morbidity and mortality in Hong Kong. To decrease the burden of diabetes and its complications on patients and society, early screening of hyperglycaemia (undiagnosed diabetes and prediabetes) is needed. The Finnish Diabetes Risk Score (FINDRISC) which has been used worldwide as a hyperglycaemia screening and diabetes prediction tool is easy, convenient, non-expensive and non-invasive. However, to date FINDRISC has not been used to classify hyperglycaemia and predict diabetes in Hong Kong. Therefore, it is important to explore the concurrent and predictive validity of FINDRISC before use. This thesis aims to (i) quantify the risk of diabetes among persons with prediabetes on the basis of the current evidence, (ii) identify an appropriate FINDRISC cut-off score to classify hyperglycaemia and normoglycaemia in the Hong Kong population, and (iii) estimate the relative risk (RR) for diabetes among persons with a higher FINDRISC compared with those with a lower score. First, a meta-analysis was performed to examine the RR of diabetes among prediabetic patients compared with normoglycaemic individuals using existing cohort studies. To explore the heterogeneity in the estimates, meta-regression was performed. Then, a cross-sectional study was conducted to determine the cut-off score of FINDRISC in identifying hyperglycaemia in the Hong Kong population. Hyperglycaemic status was tested by HbA1c tests. Receiver operating characteristic curve analysis was used to test the sensitivity and specificity of the cut-off scores. Finally, a longitudinal survey was conducted to estimate the risk of developing diabetes over one year according to FINDRISC. Logistic regression was used to calculate the unadjusted and adjusted odds ratios (OR) for diabetes incidence using FINDRISC at its own value and the high and low risk groups categorised according to the cut-off score of FINDRISC. From the meta-analysis of 36 studies, the pooled estimated RR for diabetes among all types of prediabetes as compared with the normoglycaemic population was 6.42 (95% CI: 5.26 to 7.83), and only follow-up duration (exp(β)=0.91, 95% CI: 0.84 to 0.98, p=0.012) had a significant impact on RR. To identify hyperglycaemia by using FINDRISC, the optimal cut-off point was 9, with an area under the receiver operating characteristic (ROC) curve of 0.67 (p<0.01, 95% CI: 0.60-0.74), sensitivity of 0.70 (95% CI: 0.58-0.80) and specificity of 0.57 (95% CI: 0.47-0.66). After adjusting for sex and educational level, people in the high risk group had higher odds of developing diabetes (OR: 4.59, 95% CI: 1.01 to 20.81, p= 0.048). The recommended cut-off point of FINDRISC could help to identify hyperglycaemia and predict diabetes among the Hong Kong population. The mobile application adopting this risk score can be promoted in the Hong Kong population to support diabetes self-assessment and screening. With the early screening of hyperglycaemia, early lifestyle modification and prevention of complications could be implemented, which could help patients to improve their quality of life and prognosis.