Development and Validation of DM and Pre-DM Risk Prediction Model

Diabetes Mellitus (DM) is the second most common chronic non-communicable disease (NCD) and a major public health issue. In 2017, it was estimated that 451 million adults worldwide had DM, a number that is anticipated to rise to 693 million by 2045. In terms of economic burden, it was estimated that the global cost of DM in 2015 was 1·31 trillion US dollars, which accounted for 1.8% of global gross domestic product. In China, the prevalence of DM has increased rapidly, from less than 1% in 1980 to 10.9% in 2013, with approximately 109.6 million Chinese adults (25.8% of all cases worldwide) currently living with the condition. Among the Chinese population, Hong Kong has one of the highest prevalence of DM. The Population Health Survey (PHS) 2014/2015 conducted by Department of Health found a prevalence of 8.4% of DM among persons aged 15-84 in Hong Kong, more than half (4.5%) of which were previously unknown. Data (unpublished) from the Population Health Survey 2014/2015 showed a further 9.5% of persons aged 15-84 had hyperglycaemia (pre-diabetes) but were unaware of the problem before the survey DM can result in severe complications, which lead to disabling morbidity and premature mortality. A number of randomised controlled trials (RCTs) have found that lifestyle interventions (e.g., diet, exercise) and pharmacological treatments are effective in preventing DM and its complications. However, it has been reported that 224 million adults (49.7% of all cases) world-wide are unaware that they have the condition, similar to the finding of the Hong Kong PHS 2014/2015. DM can be present for 9-12 years prior to a diagnosis and is often only detected when patients present with complications. Hence, there is an urgent need for earlier detection of DM so that appropriate interventions can be provided to prevent and/or delay progression to complications. It would be even more effective if individuals could be identified at the pre-diabetes (pre-DM) stage when there may still be an opportunity to revert to normoglycaemia by life-style modifications. While DM satisfies all Wilson and Jungner's (1968) criteria of screening studies have shown that general population screening is not effective and the current recommendation is case finding targeting at high-risk individuals. The Hong Kong Reference Framework for Diabetes Care for Adults in Primary Care Settings recommends periodic screening for DM among persons aged >=45 years old or having DM risk factors. The recommended methods for screening include 75-g oral glucose tolerance tests (OGTT), fasting plasma glucose (FPG) tests or HbA1c tests. Indeed, a cost-effectiveness analysis reported that screening for DM and prediabetes was cost-saving among patients identified as being at high risk (e.g., body mass index (BMI) > 35 kg/m(2), systolic blood pressure ≥ 130mmHg or > 55 years of age) when compared with no screening. In order to identify high risk individuals more accurately, multivariate risk prediction models have been developed and incorporated into DM prevention programs in a number of Western countries. Such models have included sociodemographic factors (e.g., age, sex), clinical factors (e.g., family history of DM, gestational DM) or biomarkers (e.g., BMI, blood pressure). However, the majority of these models were developed primarily in Caucasian populations and have not performed as well among Chinese populations. For example, the Prospective Cardiovascular Münster, Cambridge, San Antonia and Framingham models were found to have inferior discrimination in a cohort of Chinese people. This can be due to ethnic differences as well as lifestyle and socioeconomic factors, calling for the need of population-specific risk prediction models. Since 2009, a number of risk prediction models and scoring algorithms have been developed specifically for Chinese populations, mostly from Mainland China, only two of which were developed and validated on Hong Kong Chinese people. The first used simple self-reported factors and laboratory measurements to develop a scoring algorithm. However, the generalisability of the model to primary care patients may be limited as 70% of the subjects of the development and validation samples had known risk factors for DM. The second risk prediction model for Hong Kong Chinese was previously developed by members of the investigators' team with data from 3,357 asymptomatic non-diabetic professional drivers. Non-laboratory risk factors included age, BMI, family history of DM, regular physical activity (PA), and high blood pressure. Triglyceride was added to the laboratory-based algorithm. The application of this risk predication model is limited because the sample was predominately male (92.7%) professional drivers and the accuracy was modest. It is noted that the majority of factors included in previous models are non-modifiable (e.g., family history of DM, gestational DM, age), and there is a call for future research to incorporate more lifestyle factors in order to improve the predictive validity and impact of risk prediction models. Lifestyle factors that may be associated with DM and pre-DM include physical activity (PA) level, dietary factors (e.g., fibre, sugar or fat intake), alcohol consumption, smoking and sleep. This proposed study aims to develop a new DM and pre-DM risk prediction model specific for the Hong Kong general Chinese population that incorporates traditional and modifiable life style factors. The investigators will apply the novel method of machine learning as well as the traditional logistic regression in model development to improve predictive power. The investigators hope the results will enable the implementation of effective case finding of DM and pre-DM in primary care, and prevent mortality and morbidity from this common but silent NCD for the people in Hong Kong.