A decade of collaboration between CUHK and Oxford University leads to the development of the first Chinese Diabetes Outcome Model
(From left) Professor Juliana Chan Chung-ngor, from the Department of Medicine and Therapeutics at CU Medicine and Director of the Hong Kong Institute of Diabetes and Obesity; Professor Philip Clarke, Director of the Health Economics Research Centre at the University of Oxford; Professor Philip Chiu Wai-yan, Dean of CU Medicine; Professor Andrea Luk On-yan, Associate Professor and health economist Professor Juliana Lui Nga-man and first author of the study and postdoctoral fellow Dr Eric Lau Siu-him from the Department of Medicine and Therapeutics at CU Medicine.
To assess the effectiveness of CDOM in predicting outcomes in patients with type 2 diabetes, the CU Medicine research team used data from the HKDR and an independent cohort of 170,000 patients, both with more than seven years of follow-up, for dual validation. Results showed the predictions are highly consistent with actual observation of patient outcomes, making the CDOM a valuable tool for predicting long-term diabetes progression and the associated healthcare burden. The CDOM can also be used to assess the effectiveness and cost-effectiveness of existing and novel interventions, as well as the overall expenditure, thereby informing policy decisions in diabetes management and effective resource allocation. The findings have been published in the international journal Diabetes Care.
CUHK develops Chinese-specific model inspired by Oxford’s research methodology
Disease progression varies among patients with type 2 diabetes depending on their health condition, risk factors and how well they manage their blood glucose level and other risk factors. Using data from the United Kingdom Prospective Diabetes Study (UKPDS), the University of Oxford built the first diabetes outcome model and demonstrated how changing risk factors can impact the future development of these devastating complications. By estimating the effects of different interventions on risk factors and their comparative cost-effectiveness, the UKPDS outcome model has been used to inform decision-making on procurement and reimbursement policies for new technologies that can be used to treat diabetes.
Professor Philip Clarke
Professor Philip Clarke, Director of the Health Economics Research Centre at the University of Oxford and the chief developer of the UKPDS outcome model, said: “While some of these risk factors are common to both European and Asian populations, inter-ethnic differences in risk profiles, disease trajectories and complications make population-specific models important to inform context-relevant decisions, taking into consideration local healthcare financing policies, healthcare systems and users’ perspectives. As an example, Asian patients are more prone to develop diabetes at a younger age and lower body weight, more likely to develop stroke and kidney failure, and more responsive to certain treatments. These east-west differences will influence policy and treatment choice priorities. The long-term follow-up of the HKDR with detailed documentation of risk factors and outcomes has made CDOM a very useful tool, one that is highly relevant to Chinese people and to other Asian populations who share similar risk patterns.”
Among every 100 patients with diabetes, 18 develop major complications each year
The CU Medicine team used data from the HKDR between 2002 and 2019 to track over 21,000 patients with type 2 diabetes for eight years. Dr Eric Lau Siu-him, first author of the study and postdoctoral fellow in the Department of Medicine and Therapeutics at CU Medicine, used routine measures such as age, sex, disease duration, body weight, blood pressure, lipids, HbA1c (average value of blood glucose over past three months), kidney function and past history of events to develop a series of risk equations aimed at predicting changes in risk factors and diabetes complications on a yearly basis. The model was validated in an independent cohort of 176,210 patients with type 2 diabetes, followed up for seven years, with high concordance between the predicted and observed event rates.
The results showed the incidence of any major event was 18 per 100 patient-years. Apart from cancer, the CDOM demonstrated satisfactory accuracy in predicting the risk of major complications in Hong Kong Chinese patients with type 2 diabetes. The concordance indices between the estimated and observed cumulative incidence rates for most events exceeded 0.7, ranging from 0.74 (hemorrhagic stroke) to 0.94 (end-stage kidney disease), indicating the high performance and utility of CDOM for application. (Please refer to Figure 1 for details)
Figure 1: Comparison of the cumulative incidence of complications predicted by the Chinese Diabetes Outcome Model and their observed incidence
Professor Andrea Luk
Professor Andrea Luk On-yan from the Department of Medicine and Therapeutics at CU Medicine said: “Because each complication raises the likelihood of subsequent events, the variables in the risk equations are constantly updated by incorporating predicted new events to forecast recurrent events and additional complications. This huge amount of data can give rise to almost infinite permutations and combinations of risk factors and complications, resulting in a diversity of outcomes.”
Healthcare expenses for patients with diabetes increase five to 17 times after complications
Using data from the HKDR, CU Medicine researchers also reported measures of an impaired quality of life (QoL) index in over 19,000 patients with diabetes. The patients scored an average of 0.8 on the index, with 1 indicating optimal health. The top three conditions with the greatest decline in QoL were all cardiovascular events: hemorrhagic stroke (-0.230), ischemic stroke (-0.164) and peripheral vascular disease (-0.117). When any one of these 10 events occurred, the QoL index dropped, with the lower number persisting and never returning to the baseline value.
They also evaluated the healthcare costs of these 10 major events, which was five to 17 times higher than that before the events occurred. These high treatment costs were mainly caused by hospitalisations, including admissions to intensive care units, hospital wards, convalescent hospitals and infirmaries in post-event years.
Professor Juliana Luk
Professor Juliana Lui Nga-man, Assistant Professor and health economist in the Department of Medicine and Therapeutics at CU Medicine, said: “The complications with the highest healthcare costs during the year of occurrence were lower limb amputation (US$31,302), hemorrhagic stroke (US$21,164), ischemic stroke (US$17,976) and end-stage kidney disease (US$14,774). In the post-event years, end-stage kidney disease, lower limb amputation, hemorrhagic stroke and new-onset cancer incurred the highest healthcare costs, and they persisted, never returning to baseline levels. By integrating these QoL measures and healthcare costs with the CDOM, we can estimate the impact of modifying different risk factors, such as blood pressure, blood lipids, blood glucose and body weight, and of the use of organ protective drugs, on the risk of future events. We can also compare the cost-effectiveness of new versus old treatments on these expensive-to-treat events in different patient groups and on an individual level.”
Professor Juliana Chan
Professor Juliana Chan Chung-ngor from the Department of Medicine and Therapeutics at CU Medicine, Director of the Hong Kong Institute of Diabetes and Obesity and the senior investigator of the HKDR and CDOM, said: “Every person has a unique set of modifiable and non-modifiable risk factors, which interact in a most complex manner to give rise to different complications in people with diabetes. By creating different scenarios, a doctor can have a more engaging discussion with his or her patient on how to individualise control of risk factors and select medications to help him or her stay healthy and live a long life. These models also put whoever is responsible for paying for these treatments, whether public or private, in a better position to allocate resources or design payment plans, making healthcare more accessible, sustainable and affordable.”
Professor Chan mentioned an example of a 53-year-old man diagnosed with diabetes 10 years before, who is obese, smokes and has poor control of blood pressure, blood lipids and blood glucose, and has a history of ischaemic heart disease. He is projected to live for 12 more years at most, preceded by multiple admissions due to complications. By quitting smoking, losing weight and gaining better control of risk factors, his life expectancy can be extended to 18 years, with considerable savings in healthcare costs and human suffering.
CU Medicine and the University of Oxford are entering a strategic partnership to raise awareness, build capacity and conduct collaborative projects in health economics and health technology assessment. The partnership supports the vision of the Government of the Hong Kong Special Administrative Region of the People’s Republic of China of integrating healthcare delivery, clinical trials and the health industry, and of turning Hong Kong into a hub of innovative medicine.
This study was funded by the Health and Medical Research Fund and Public Policy Research Funding Scheme from the Food and Health Bureau of the Government of the HKSAR, and the Asia Diabetes Foundation.
Data on healthcare costs and quality of life in 19,800 patients with type 2 diabetes enrolled on the HKDR is available online:
