Cover story for Pulse@UM Year 2021 Issue 1: Big Data & Artificial Intelligence in Medicine and Healthcare Research

Big Data in Diabetes Care: Advancing Quality and Improving Outcomes

Blue vector created by vectorjuice - www.freepik.com
Blue vector created by vectorjuice – www.freepik.com

By Associate Professor Dr Lim Lee Ling, Dr Lim Quan Hziung & Dr Nurul Aida Binti Mohamad Zulkafli

Big Data, put simply, is the analysis of information, not only at massive volumes but of great variety and high velocity, in order to yield outcomes of substantial veracity and values. In 2012, the International Data Corporation
projected the value of the digital universe to reach a mind-boggling of 40,000 Exabytes (40 × 1018 bytes) by the year 2020. Leveraging on this, Big Data has been utilised by industries worldwide for different purposes. Advancements in data science have enabled the development of machine learning and artificial intelligence that can facilitate health predictions.

Figure 1: Workflow of Big Data analytics.

In health care, Big Data is implemented by digitally consolidating different parameters of health information (Figure 1). This information is gathered from complex and heterogeneous platforms, including but is not limited to, electronic health records, prescription information, geospatial mobility patterns, and data from clinical trials, laboratory, imaging, digital health technology, and multi-omics analytics. This massive volume of data is compiled in a data warehouse or registry, and subsequently processed via data wrangling in order to deliver key metrics to multi- level stakeholders for health services planning, improving health outcomes and reducing health care costs in the long run.

The care of people with diabetes stands to benefit the most from the implementation of data science. In addition to genetic predisposition, rapid globalisation, urbanisation, and lifestyle transitions have accelerated the diabetes epidemic in Asian countries including Malaysia. The understanding of current diagnostic, prognostic, and therapeutic areas in diabetes has undergone evolutionary changes over the years.

To complement the findings of randomised clinical trials, there are growing demands for high-quality real-world evidence to provide insights into diabetes epidemiology, treatment effectiveness and safety, and health economic impact. Integrating population- and individual-level data on preventive medicine; systems biology, information technology, and ecological models will enhance the implementation of precision medicine in diabetes (Figure 2). This data-driven approach is necessary to guide policy, treatment and reimbursement decision-making.

Figure 2: Transforming fragmented care into a multi-component integrated care system to support the delivery of high-quality diabetes care.

In the current unprecedented COVID-19 pandemic, real-world evidence through robust data analytics has shown that people with diabetes have 1.4 to 2.4 times increased risk of death or intensive care admission, suggesting
fundamental issues in the organisation and delivery of diabetes care. Given the complex care needs of people with diabetes, a multi-level and multi-component strategy is crucial to improve diabetes prevention and management (Figure 2 and Figure 3). Several high-income countries and regions including the USA, United Kingdom, Hong Kong and Singapore have successfully leveraged data to drive actions at policy, system, provider and individual levels for tackling the syndemic of type 2 diabetesand obesity (Figure 3). This nation- and territory-wide initiatives,include but are not limited to, diet-related policies (e.g., food labelling regulations, sugar-sweetened beverage taxes, healthy school food), diabetes prevention programmes, risk assessment and management programmes, and patient support programmes.

Figure 3: A conceptual framework for a multi-level and multi-component society–community–individual strategy to integrate primary and secondary prevention of diabetes.

Closer to home, there are ongoing efforts to utilise data to improve disease surveillance including the setting up of MeLODY (Multi-ethnic Lifestyle, Obesity, and Diabetes Registry in MalaYsia; contact limleeling@um.edu.
my) and MeMORY (Multi-ethnic Maternal Offspring Registry in MalaYsia; contact shireene.vethakkan@gmail.com). These approaches are timely in Malaysia, given the unabated rise of diabetes prevalence to 18.3% in 2019.

To this end, Big Data is a revolutionary tool to facilitate precision medicine in diabetes. Multi-level stakeholders including policymakers, payers, communities, health care providers, and people with diabetes must join efforts to unlock key discoveries and outline the opportunities for winning the battle against diabetes.


Other posts: